**Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies**

Claudine Mélan and Edith Galy

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/50342

#### **1. Introduction**

34 Advances in Air Navigation Services

p. 21, AIAA-2010-7542.

Cybernatics, Vol. 1, pp. 802–808.

49th Annual Meeting, pp. 422-426.

Journal of Industrial Engineering, 16(1), pp. 61-70.

Aviation Psychology, Dayton, Ohio, USA.

Portland, Oregon.

AIAA-2010-7541.

Technology

Technology.

Operations Forum, Los Angeles, CA. -31

An Air Traffic Management Metric. NASA-TM-1998-112226.

[20] Masalonis, A.J., Calaham, M.B. and Wanke, C.R. (2003). Dynamic Density and Complexity Metrics for Realtime Traffic Flow Management. The MITRE Corp. McLean, VA. [21] Chatterji, G.B. & Sridhar, B. (2001). Measures for Air Traffic Controller Workload Prediction. Proceedings of the First AIAA Aircraft Technology, Integration, and

[22] Laudeman, I.V., S.G. Shelden, R. Branstrom, & C.L. Brasil, (1999). Dynamic Density:

[23] d'Engelbronner, J., Mulder, M., van Paassen, M. M., de Stigter, S., and Huisman, H. (2010). The Use of the Dynamic Solution Space to Assess Air Traffic Controller Workload, AIAA Guidance, Navigation, and Control Conference, AIAA, Toronto, CA,

[24] Abdul Rahman S. M. B., Mulder M. and van Paassen M. M. (2011). Using the Solution Space Diagram in Measuring the Effect of Sector Complexity During Merging Scenarios, Proceeding of AIAA Guidance, Navigation, and Control Conference,

[25] Van Dam, S. B. J. V., Abeloos, A. L. M., Mulder, M., and van Paassen, M. M. (2004). Functional Presentation of Travel Opportunities in Flexible Use Airspace: an EID of an Airborne Conflict Support Tool, IEEE International Conference on Systems, Man and

[26] Mercado-Velasco, G., Mulder, M., and van Paassen, M. M. (2010). Analysis of Air Traffic Controller Workload Reduction Based on the Solution Space for the Merging Task, AIAA Guidance, Navigation, and Control Conference, AIAA, Toronto, CA, p. 18,

[27] Rantanen, E. M. and Nunes, A. (2005): Hierarchical Conflict Detection in Air Traffic

[28] Remington, R. W., Johnston, J. C., Ruthruff, E., Gold, M., Romera. M. (2000). Visual Search in Complex Displays: Factors Affecting Conflict Detection by Air Traffic

[29] Nunes, A. and Kirlik, A. (2005). An Empirical Study of Calibration in Air Traffic Control Expert Judgment, Proceedings of the Human Factors and Ergonomics Society

[30] Koperdekar, P., Schwarzt, A., Magyarits, S. and Rhodes, J. (2009). Airspace Complexity Measurement: An Air Traffic Control Simulation Analysis. International

[31] Zhou, W. (2011). The 3D Solution Space: Metric to Assess Workload in Air Traffic Control. Master's Thesis. Department of Control and Simulation. Delft University of

[32] Lodder, J., Comans, J., van Paassen, M. M. and Mulder, M. (2011). Altitude-extended Solution Space Diagram for Air Traffic Controllers. International Symposium on

[33] Tattersall, A. and Foord, P. (1996), An Experimental Evaluation of Instantaneous Self-Assessment as a Measure of Workload, Ergonomics, Vol. 39, No. 5, pp. 740–748. [34] d'Engelbronner, J. G. (2009). Construction of a Tangent-Based Solution Space Diagram. Unpublished MSc. Thesis. Faculty of Aerospace Engineering, Delft University of

Control, The International Journal of Aviation Psychology, 15:4, 339-362

Controllers, Human Factors, Vol. 42, No. 3, Fall 2000, pg 349-366.

Air traffic controllers' (ATCs') work evolves constantly, concerning in particular route complexity and traffic density, but also development of supporting technology. Introducing more automation to allow more efficient ATC control and increased safety and security also requires enhanced supervisory activity, situation awareness, processing of larger amounts of data. These cognitive processes place a heavy load on ATCs' memory functions as they require item processing and recall, which are also involved in control operations such as monitoring traffic, controlling aircraft movements, managing air traffic sequences, resolving conflicts. Better understanding of memory processes and of their limitations in expert ATCs may thus be crucial for the development of future automation tools, but also for training and selection of controllers. The aim of the present contribution is to give a comprehensive overview of memorisation performance in air traffic controllers, in light of the most recent memory models. More especially, a series of experiments reveal that ATCs' memorisation performance varies in a complex manner according to both task-related factors (presentation modality, number of items, recall protocol), and task-independent factors. The latter are related more especially to shift-scheduling (time-of-day, on-shift time) and physiological capacities (alertness, automatic item processing).

#### **2. ATCs' performance variations according to task-related factors**

#### **2.1. Information processing during control operations**

En route ATC involves the processing of information relative to a variable number of aircrafts coming from different directions, at diverse speed and altitudes, and heading to

#### 36 Advances in Air Navigation Services

various destinations or that are, on contrary, grouped in a more restricted space, thus requiring more in-depth processing in order to anticipate air-plane conflicts. Presentation modality is of interest as information about an aircraft is presented visually on a strip or script 10 to 15 min before its real-time presentation in the visual modality (radar information) or in the auditory modality (radio information). Radar information includes instantaneous level, attitude (stable, climbing, descending) and speed group, and strips present many information as aircraft call sign, aircraft type and associated speed (or power), provenance and destination, route case (estimated hours of flying), and three cases completed by ATCs with information concerning coordination information with other ATCs, or changes requested in flight.

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 37

*2.2.1. Differences during processing of auditory and visually presented material* 

memory traces and ensure higher recall for heard items (Frankish, 1985, 2008).

*2.2.2. Differences during restitution of heard and seen item lists* 

sounds.

Several authors suggested that a long-lasting sensory acoustic trace would be generated for auditory presented words but not for visually presented words (Cowan, 1984; Crowder & Morton, 1969; Crowder et al., 2004; Penney, 1989). The term acoustic–sensory information refers to sensory representations of sounds. Though there is some evidence of a sensory visual trace for seen words, it would be very short-lived compared to the long-lasting sensory acoustic trace that would favour more efficient encoding of heard material. In addition, orally presented list-items would be associated with temporal cues that would not be generated by successively presented visual words. These temporal cues would then result in stronger

Other authors argued that the differences between the visual and the auditory information processing streams would occur when the sensory trace is processed into a short term memory trace. The working memory model (Baddeley & Hitch, 1974) includes a dedicated phonological subsystem, in which the code used to represent verbal items broadly corresponds to the phonetic level. Hence, auditory information would benefit from automatic phonological coding while visual information would require effortful phonological recoding (Baddeley, 1986, 2000; Penney, 1989). Phonological coding would favour mental rehearsal and maintenance of information in a short-term store (Baddeley, 1986), what would result in a larger number of phonologically coded memory traces in a short-term store following hearing than following seeing item-lists. According to Penney (1989) the auditory superiority would be based on the combination of a longer-lasting sensory acoustic trace and automatic phonological coding of orally presented materiel.

Differences during item recall were also proposed to account for better immediate recall of heard than of seen materiel. This research is based in particular on the robust finding that when participants recall a sequence of spoken digits the last one is almost always correctly recalled, but if the same sequence is presented visually, recall of the final item is relatively poor. According to Crowder and Morton (1969), this auditory recency effect indicates that an acoustically coded representation of the final list item is maintained in a sensory store, while representations of earlier items are overwritten by successively incoming speech

Alternatively, Cowan and co-workers accounted for the recency effect by suggesting greater resistance to output interference for heard compared to seen verbal material (Cowan et al., 2002; Harveay & Beaman, 2007; Madigan, 1971). Output interference is defined as the degradation of memory representations as recall proceeds across output positions. Evidence in favour of a greater resistance to interference for heard items came from studies reporting a marked auditory superiority when restitution was based on a free recall procedure (without providing any cues) and a reduced effect when a recognition procedure was used (list items are presented together with new items). Item recognition may be explained by an impression of familiarity for the list-items and would thus not require in-depth processing.

A number of simulation studies and *in situ* ATC observations explored the way such factors may impact on memory performance. Thus, expert ATCs show high recall performance of aircrafts and their position on a sector map, and poorer recollection of details regarding an aircraft (Means et al., 1988). Higher recognition accuracy for aircrafts involved in an impending conflict compared to those that would not cross or cross in some near future confirmed the impact of aircraft status on memory performance (Gronlund et al., 1998, 2005). Simulation studies revealed higher recall when navigational messages were presented in the auditory rather than in the visual modality. In addition, performance dropped considerably when message length increased beyond three commands, while command wordiness (2 or 4 words) had only limited effects on recall (Wickens & Hollands, 2000; Barshi & Healy, 2002; Schneider et al., 2004).

#### **2.2. Information processing according to presentation modality: The auditory superiority effect**

Controlled laboratory studies have systematically explored task-related factors that may affect mnemonic performance and have led to the proposal of integrated theoretical models. More especially, encoding and processing of auditory compared to visual verbal material has systematically revealed superior recall of heard material in short-term memory (for a review, Penney, 1989). A common test in this research field contrasts silent reading and reading aloud of unrelated lists of words, nonwords, letters, or digits (see, e.g., Conrad & Hull, 1968; Frankish, 1985). It is assumed that both silent reading and reading aloud provide phonological information and that only reading aloud provides acoustic–sensory information. This effect was found for the recency part of the serial position curve in immediate free recall and in serial recall. Overt vocalization of a visually presented list by the subject produced much the same effect as did auditory presentation on the recency part of the serial position curve, but subject vocalization tended to reduce recall in the non recency part of the serial position curve. The modality effect has been observed with both written and oral recall, but seems to be slightly more marked in the former case. This auditory superiority effect, known as the modality effect, was accounted for by modalityrelated processing differences at different stages of the memorisation process, rather than by strategic differences.

#### *2.2.1. Differences during processing of auditory and visually presented material*

36 Advances in Air Navigation Services

ATCs, or changes requested in flight.

Barshi & Healy, 2002; Schneider et al., 2004).

**superiority effect** 

strategic differences.

various destinations or that are, on contrary, grouped in a more restricted space, thus requiring more in-depth processing in order to anticipate air-plane conflicts. Presentation modality is of interest as information about an aircraft is presented visually on a strip or script 10 to 15 min before its real-time presentation in the visual modality (radar information) or in the auditory modality (radio information). Radar information includes instantaneous level, attitude (stable, climbing, descending) and speed group, and strips present many information as aircraft call sign, aircraft type and associated speed (or power), provenance and destination, route case (estimated hours of flying), and three cases completed by ATCs with information concerning coordination information with other

A number of simulation studies and *in situ* ATC observations explored the way such factors may impact on memory performance. Thus, expert ATCs show high recall performance of aircrafts and their position on a sector map, and poorer recollection of details regarding an aircraft (Means et al., 1988). Higher recognition accuracy for aircrafts involved in an impending conflict compared to those that would not cross or cross in some near future confirmed the impact of aircraft status on memory performance (Gronlund et al., 1998, 2005). Simulation studies revealed higher recall when navigational messages were presented in the auditory rather than in the visual modality. In addition, performance dropped considerably when message length increased beyond three commands, while command wordiness (2 or 4 words) had only limited effects on recall (Wickens & Hollands, 2000;

**2.2. Information processing according to presentation modality: The auditory** 

Controlled laboratory studies have systematically explored task-related factors that may affect mnemonic performance and have led to the proposal of integrated theoretical models. More especially, encoding and processing of auditory compared to visual verbal material has systematically revealed superior recall of heard material in short-term memory (for a review, Penney, 1989). A common test in this research field contrasts silent reading and reading aloud of unrelated lists of words, nonwords, letters, or digits (see, e.g., Conrad & Hull, 1968; Frankish, 1985). It is assumed that both silent reading and reading aloud provide phonological information and that only reading aloud provides acoustic–sensory information. This effect was found for the recency part of the serial position curve in immediate free recall and in serial recall. Overt vocalization of a visually presented list by the subject produced much the same effect as did auditory presentation on the recency part of the serial position curve, but subject vocalization tended to reduce recall in the non recency part of the serial position curve. The modality effect has been observed with both written and oral recall, but seems to be slightly more marked in the former case. This auditory superiority effect, known as the modality effect, was accounted for by modalityrelated processing differences at different stages of the memorisation process, rather than by Several authors suggested that a long-lasting sensory acoustic trace would be generated for auditory presented words but not for visually presented words (Cowan, 1984; Crowder & Morton, 1969; Crowder et al., 2004; Penney, 1989). The term acoustic–sensory information refers to sensory representations of sounds. Though there is some evidence of a sensory visual trace for seen words, it would be very short-lived compared to the long-lasting sensory acoustic trace that would favour more efficient encoding of heard material. In addition, orally presented list-items would be associated with temporal cues that would not be generated by successively presented visual words. These temporal cues would then result in stronger memory traces and ensure higher recall for heard items (Frankish, 1985, 2008).

Other authors argued that the differences between the visual and the auditory information processing streams would occur when the sensory trace is processed into a short term memory trace. The working memory model (Baddeley & Hitch, 1974) includes a dedicated phonological subsystem, in which the code used to represent verbal items broadly corresponds to the phonetic level. Hence, auditory information would benefit from automatic phonological coding while visual information would require effortful phonological recoding (Baddeley, 1986, 2000; Penney, 1989). Phonological coding would favour mental rehearsal and maintenance of information in a short-term store (Baddeley, 1986), what would result in a larger number of phonologically coded memory traces in a short-term store following hearing than following seeing item-lists. According to Penney (1989) the auditory superiority would be based on the combination of a longer-lasting sensory acoustic trace and automatic phonological coding of orally presented materiel.

#### *2.2.2. Differences during restitution of heard and seen item lists*

Differences during item recall were also proposed to account for better immediate recall of heard than of seen materiel. This research is based in particular on the robust finding that when participants recall a sequence of spoken digits the last one is almost always correctly recalled, but if the same sequence is presented visually, recall of the final item is relatively poor. According to Crowder and Morton (1969), this auditory recency effect indicates that an acoustically coded representation of the final list item is maintained in a sensory store, while representations of earlier items are overwritten by successively incoming speech sounds.

Alternatively, Cowan and co-workers accounted for the recency effect by suggesting greater resistance to output interference for heard compared to seen verbal material (Cowan et al., 2002; Harveay & Beaman, 2007; Madigan, 1971). Output interference is defined as the degradation of memory representations as recall proceeds across output positions. Evidence in favour of a greater resistance to interference for heard items came from studies reporting a marked auditory superiority when restitution was based on a free recall procedure (without providing any cues) and a reduced effect when a recognition procedure was used (list items are presented together with new items). Item recognition may be explained by an impression of familiarity for the list-items and would thus not require in-depth processing.

#### 38 Advances in Air Navigation Services

In contrast, free recall requires subjects to explicitly recall each item, thereby degrading the remaining memory traces (Brébion et al., 2005).

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 39

> 34 13 2

> 30 4 2

16 3 1

14 1 1

Visual Auditory

**Table 1.** Number of each of the terminal item sequences as the opening run (upper panel) and in noninitial response positions (lower panel) during free recall of 6- and 9-word lists with auditory and visual

As shown in table 1, ATCs recalled complete 6-item lists three times more frequently following auditory (27) than following visual list (10) presentation, while recall of five-item sub-sequences (2-3-4-5-6) was rare in both modalities (2 vs. 7 respectively). These results further indicate that occurrences for heard lists were tenfold more frequent for complete sixitem lists than for five-item sub-sequences, while no such difference occurred for seen lists. Occurrences of equivalent end sub-sequences of 9-item lists (5-6-7-8-9; 4-5-6-7-8-9) ranged between zero and three in both modalities. The auditory recall advantage of these lists appeared to result from higher occurrences of ordered 2- and 3-item end sub-sequences in

Taken together, the findings show that ATCs would spontaneously adopt the output strategy consisting of uttering end sub-sequences more frequently for heard than for seen items, leading to, or contributing to significant higher overall performance. They thus further stress the proposal that "auditory presentation seems to protect the end of the list from output interference" (Cowan et al., 2002, p.168). In favour of this hypothesis, these authors showed that the auditory advantage is even more pronounced when output interference is high. Alternatively it has been proposed that the auditory advantage that extends over the last few serial positions is retrieved independently for each item from an echoic trace (Frankish, 2008). According to this author, pronounced recency in immediate serial recall is limited to stimuli that engage the perceptual mechanism involved in linguistic decoding of speech. Further research intended to disentangle different sources of sensory information at input. Thus, Rummer and Schweppe (2005) observed a modality effect for spoken sentences compared to conditions without acoustic–sensory information, i.e., both silent reading and mouthing. As the latter two conditions did not differ from each other, the results would rule out articulatory information at input as a source of the modality effect. Differences due to output modality were also investigated. For written recall the auditory advantage was larger with high than with low output interference, while this difference was not maintained for spoken recall (Harvey & Beaman, 2007). Taken together, the data suggest that both superior auditory encoding and reduced output interference would contribute to

Terminal item sequences 6-word lists

Terminal item sequences 9-word lists

5, 6 4, 5, 6 3, 4, 5, 6

8, 9 7, 8, 9 6, 7, 8, 9

presentation (Galy et al., 2010).

*other than initial recall positions*.

the auditory modality effect.

Lower output interference has also been proposed to account for the auditory advantage of sentence recall (Rummer & Schweppe, 2005). In line with this hypothesis, Beaman and Morton (2000) showed that following presentation of 16-item lists, subjects preferentially recalled 2-, 3- and 4-item sub-sequences (of the form 15-16, 14-15-16, 13-14-15-16) from the end of the lists. However, while end sub-sequences were recalled with a similar frequency in both modalities on the opening run of a trial, they were recalled significantly more often for heard than for seen items during the course of a trial.

#### **2.3. Auditory superiority for heard versus seen item-lists in ATCs**

In light of these findings we explored whether output interference while recalling seen items may be reduced in participants well-practiced in processing visual information, i.e. ATCs habitually processing visual information on a radar. In addition we tested to what extent the effect may be explained by memory load, typically explored by presenting a list of items of variable length. Previous research with this kind of procedure has shown that the amount of proactive interference is much less for smaller set sizes (Oberauer & Vockenberg, 2009) although there can be some proactive interference even at small set sizes (Carroll et al., 2010). We used the procedure described by Beaman and Morton (2000) to explore free recall of heard versus seen 6- and 9-item lists in 15 volunteer ATCs of an en-route centre in southern France. Participants were aged 31.3 years (range: 27 to 42 years old) and had been working for 7 years and 4 months (range: 3 to18 years) in the control centre. The sub-sequences recalled by ATCs at initial and non-initial output positions are summarized in table 1. Analyses revealed significant higher mean numbers of items recalled following hearing (4.9 and 3.6 respectively for 6- and 9- item lists) than seeing the lists (4.2 and 3.2 respectively).



38 Advances in Air Navigation Services

6 5, 6 4, 5, 6 3, 4, 5, 6 2, 3, 4, 5, 6 1, 2, 3, 4, 5, 6

9 8, 9 7, 8, 9 6, 7, 8, 9 5, 6, 7, 8, 9 4, 5, 6, 7, 8, 9

remaining memory traces (Brébion et al., 2005).

heard than for seen items during the course of a trial.

**2.3. Auditory superiority for heard versus seen item-lists in ATCs** 

for 6- and 9- item lists) than seeing the lists (4.2 and 3.2 respectively).

Terminal item sequences 6-word lists

Terminal item sequences 9-word lists

Visual Auditory

In contrast, free recall requires subjects to explicitly recall each item, thereby degrading the

Lower output interference has also been proposed to account for the auditory advantage of sentence recall (Rummer & Schweppe, 2005). In line with this hypothesis, Beaman and Morton (2000) showed that following presentation of 16-item lists, subjects preferentially recalled 2-, 3- and 4-item sub-sequences (of the form 15-16, 14-15-16, 13-14-15-16) from the end of the lists. However, while end sub-sequences were recalled with a similar frequency in both modalities on the opening run of a trial, they were recalled significantly more often for

In light of these findings we explored whether output interference while recalling seen items may be reduced in participants well-practiced in processing visual information, i.e. ATCs habitually processing visual information on a radar. In addition we tested to what extent the effect may be explained by memory load, typically explored by presenting a list of items of variable length. Previous research with this kind of procedure has shown that the amount of proactive interference is much less for smaller set sizes (Oberauer & Vockenberg, 2009) although there can be some proactive interference even at small set sizes (Carroll et al., 2010). We used the procedure described by Beaman and Morton (2000) to explore free recall of heard versus seen 6- and 9-item lists in 15 volunteer ATCs of an en-route centre in southern France. Participants were aged 31.3 years (range: 27 to 42 years old) and had been working for 7 years and 4 months (range: 3 to18 years) in the control centre. The sub-sequences recalled by ATCs at initial and non-initial output positions are summarized in table 1. Analyses revealed significant higher mean numbers of items recalled following hearing (4.9 and 3.6 respectively

**Table 1.** Number of each of the terminal item sequences as the opening run (upper panel) and in noninitial response positions (lower panel) during free recall of 6- and 9-word lists with auditory and visual presentation (Galy et al., 2010).

As shown in table 1, ATCs recalled complete 6-item lists three times more frequently following auditory (27) than following visual list (10) presentation, while recall of five-item sub-sequences (2-3-4-5-6) was rare in both modalities (2 vs. 7 respectively). These results further indicate that occurrences for heard lists were tenfold more frequent for complete sixitem lists than for five-item sub-sequences, while no such difference occurred for seen lists. Occurrences of equivalent end sub-sequences of 9-item lists (5-6-7-8-9; 4-5-6-7-8-9) ranged between zero and three in both modalities. The auditory recall advantage of these lists appeared to result from higher occurrences of ordered 2- and 3-item end sub-sequences in *other than initial recall positions*.

Taken together, the findings show that ATCs would spontaneously adopt the output strategy consisting of uttering end sub-sequences more frequently for heard than for seen items, leading to, or contributing to significant higher overall performance. They thus further stress the proposal that "auditory presentation seems to protect the end of the list from output interference" (Cowan et al., 2002, p.168). In favour of this hypothesis, these authors showed that the auditory advantage is even more pronounced when output interference is high. Alternatively it has been proposed that the auditory advantage that extends over the last few serial positions is retrieved independently for each item from an echoic trace (Frankish, 2008). According to this author, pronounced recency in immediate serial recall is limited to stimuli that engage the perceptual mechanism involved in linguistic decoding of speech. Further research intended to disentangle different sources of sensory information at input. Thus, Rummer and Schweppe (2005) observed a modality effect for spoken sentences compared to conditions without acoustic–sensory information, i.e., both silent reading and mouthing. As the latter two conditions did not differ from each other, the results would rule out articulatory information at input as a source of the modality effect. Differences due to output modality were also investigated. For written recall the auditory advantage was larger with high than with low output interference, while this difference was not maintained for spoken recall (Harvey & Beaman, 2007). Taken together, the data suggest that both superior auditory encoding and reduced output interference would contribute to the auditory modality effect.

As ATCs are well-practiced in processing successively presented visual information, on contrary to participants in the above-cited laboratory experiments, the present findings favour the idea that differential physiological features may characterize the visual and auditory information processing streams (Penney, 1989). The modality effect may then be implemented in the current models of spoken and of written words, which are divided on the relation between the different levels of processing, which will not be discussed in this issue. Briefly, in some models the transmission of information through successive levels of representation is represented as a unidirectional flow within a feed forward network. Perceptual analysis begins with encoding of acoustic features, which are then translated into phonetic and then lexical representations. In contrast, interactive activation models propose that communication between these levels is bidirectional. Whenever a lexical unit becomes active, feedback connections boost activation of the units that represent its constituent phonemes (for a review, Frankish, 2008).

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 41

occurrences of incidents on the night shift when the circadian decline in human capabilities is further aggravated by a chronic sleep deficit and fatigue (Costa, 2003; Folkard & Akersted, 2003). Laboratory studies have established that alertness is low in the morning, increases during the day until the late afternoon, before decreasing in the evening and reach a minimal level on early morning hours. The shift work literature revealed that in real-job situations alertness also varies with time of day and that the typical diurnal trend would only be marginally modified by shift work scheduling features. However, early morning shifts, extended shift duration and repeated night shifts have been shown to be associated with increased sleepiness, more especially during the last half of extended shifts, particularly on night shifts (Kecklund et al., 1997; Rosa, 1995), but also on day shifts (Tucker

Circadian variations appear to be less consistent for other psychological measures recorded in several shift work studies, and in particular for self-reported tension (Folkard, 1990; Kecklund et al., 1997; Monk et al., 1985; Owens et al., 2000 ; Prizmic et al., 1995). While some studies reported a circadian trend for perceived tension, others did not, and still others reported an atypical trend. More especially, operators supervising a satellite across 24h-day displayed significant increased self-rated tension and heart rate on the first hour of each shift, even on the night-shift, despite a lower baseline level for heart rate during the night (Cariou et al., 2008). In contrast, when the same satellite controllers rated Thayer's (1989) Activation-Deactivation checklist, their alertness level was highly correlated with their body temperature (Fig. 1), largely considered as an index of subjects' functional state. Both measures followed a typical circadian trend, indicating a strong dependency of these

Taken together with the shift-work literature, these results indicate that some subjective and objective measures (here alertness and body temperature) show a strong dependency on the endogenous regulation systems, as they display a circadian trend in different shiftwork conditions, like in controlled laboratory conditions, despite minor variations in this trend by external factors. This then implies that working during the circadian trough requires an additional effort as operators' functional state is at its lowest level. Most interestingly, other measures (here subjective tension and heart rate), which are known to display a circadian trend in controlled laboratory conditions, are much more influenced by external factors, so that the trends may vary considerably between situations (or jobsituations). Andorre and Queinnec (1998) reported a significant increase on the first shifthour for real-job performance (pages checked on a computer-screen) in operators controlling a chemical production process. In both studies the atypical trend was interpreted as indicating enhanced cognitive demands following shift take-over in jobsituations concerned with supervisory control of a dynamic process. Thus, in shift-work conditions, some psychological and physiological measures, and in particular those that are and other stress-sensitive, would be largely influenced by environmental factors, including meal-timing, task demands, time-pressure and so on, which may mask the otherwise circadian trend of these measures (Averty et al., 2004; Brookings et al., 1996;

et al., 1998).

measures on the endogenous regulation systems.

Khaleque, 1984; Rose et al., 1982).

#### **3. Shift-scheduling and ATCs' functional state**

Like in other safety-related job situations, ATC requires operators to work successive shifts; i.e. different teams work in succession to cover the whole 24h-day. As a consequence, controllers are subjected to the negative impact of shift work on biological rhythms, sleep, job performance, and psychological measures (Costa, 1999; Della Rocco & Nesthus, 2005; Dinges et al., 1997; Folkard & Tucker, 2003).

#### **3.1. Regulation mechanisms of circadian variations**

It is now largely accepted that the negative effects of shift work result from a disruption of the habitual circadian regulation of physiological and psychological measures (Costa, 2003; Siegrist, 2010). Circadian variations across the 24h-day are under control of two endogenous systems, the homeostatic system expressed by the fatigue accumulating since awakening, and the circadian system evidenced by a sinusoidal variation across the 24h period. Hence, shift work is systematically associated with a cumulative sleep deficit (homeostatic system), and a decreased amplitude of sinusoidal variations (circadian system), thereby interfering with the two powerful factors limiting human ability and, as a consequence, safety and security (Akerstedt, 1991; Akerstedt et al., 2004; Dinges et al., 1997; Folkard & Akerstedt, 1992; Tucker et al., 2006). As stated by Akerstedt (2007, p. 209) "Being exposed to the circadian low (during work/activity), extended time awake or reduced duration of sleep will impair performance".

#### **3.2. Circadian and non circadian variations of subjective and physiological measures in shift-workers**

In order to investigate a person's functional state, behavioural, physiological or subjective measures have been used. It is generally accepted that these measures are subjected to a circadian rhythmicity. Shift work effects have mostly been documented by reporting decreased self-rated alertness (Akerstedt & Gillberg, 1990; Galy et al., 2008) and increased occurrences of incidents on the night shift when the circadian decline in human capabilities is further aggravated by a chronic sleep deficit and fatigue (Costa, 2003; Folkard & Akersted, 2003). Laboratory studies have established that alertness is low in the morning, increases during the day until the late afternoon, before decreasing in the evening and reach a minimal level on early morning hours. The shift work literature revealed that in real-job situations alertness also varies with time of day and that the typical diurnal trend would only be marginally modified by shift work scheduling features. However, early morning shifts, extended shift duration and repeated night shifts have been shown to be associated with increased sleepiness, more especially during the last half of extended shifts, particularly on night shifts (Kecklund et al., 1997; Rosa, 1995), but also on day shifts (Tucker et al., 1998).

40 Advances in Air Navigation Services

phonemes (for a review, Frankish, 2008).

Dinges et al., 1997; Folkard & Tucker, 2003).

impair performance".

**measures in shift-workers** 

**3. Shift-scheduling and ATCs' functional state** 

**3.1. Regulation mechanisms of circadian variations** 

As ATCs are well-practiced in processing successively presented visual information, on contrary to participants in the above-cited laboratory experiments, the present findings favour the idea that differential physiological features may characterize the visual and auditory information processing streams (Penney, 1989). The modality effect may then be implemented in the current models of spoken and of written words, which are divided on the relation between the different levels of processing, which will not be discussed in this issue. Briefly, in some models the transmission of information through successive levels of representation is represented as a unidirectional flow within a feed forward network. Perceptual analysis begins with encoding of acoustic features, which are then translated into phonetic and then lexical representations. In contrast, interactive activation models propose that communication between these levels is bidirectional. Whenever a lexical unit becomes active, feedback connections boost activation of the units that represent its constituent

Like in other safety-related job situations, ATC requires operators to work successive shifts; i.e. different teams work in succession to cover the whole 24h-day. As a consequence, controllers are subjected to the negative impact of shift work on biological rhythms, sleep, job performance, and psychological measures (Costa, 1999; Della Rocco & Nesthus, 2005;

It is now largely accepted that the negative effects of shift work result from a disruption of the habitual circadian regulation of physiological and psychological measures (Costa, 2003; Siegrist, 2010). Circadian variations across the 24h-day are under control of two endogenous systems, the homeostatic system expressed by the fatigue accumulating since awakening, and the circadian system evidenced by a sinusoidal variation across the 24h period. Hence, shift work is systematically associated with a cumulative sleep deficit (homeostatic system), and a decreased amplitude of sinusoidal variations (circadian system), thereby interfering with the two powerful factors limiting human ability and, as a consequence, safety and security (Akerstedt, 1991; Akerstedt et al., 2004; Dinges et al., 1997; Folkard & Akerstedt, 1992; Tucker et al., 2006). As stated by Akerstedt (2007, p. 209) "Being exposed to the circadian low (during work/activity), extended time awake or reduced duration of sleep will

**3.2. Circadian and non circadian variations of subjective and physiological** 

In order to investigate a person's functional state, behavioural, physiological or subjective measures have been used. It is generally accepted that these measures are subjected to a circadian rhythmicity. Shift work effects have mostly been documented by reporting decreased self-rated alertness (Akerstedt & Gillberg, 1990; Galy et al., 2008) and increased Circadian variations appear to be less consistent for other psychological measures recorded in several shift work studies, and in particular for self-reported tension (Folkard, 1990; Kecklund et al., 1997; Monk et al., 1985; Owens et al., 2000 ; Prizmic et al., 1995). While some studies reported a circadian trend for perceived tension, others did not, and still others reported an atypical trend. More especially, operators supervising a satellite across 24h-day displayed significant increased self-rated tension and heart rate on the first hour of each shift, even on the night-shift, despite a lower baseline level for heart rate during the night (Cariou et al., 2008). In contrast, when the same satellite controllers rated Thayer's (1989) Activation-Deactivation checklist, their alertness level was highly correlated with their body temperature (Fig. 1), largely considered as an index of subjects' functional state. Both measures followed a typical circadian trend, indicating a strong dependency of these measures on the endogenous regulation systems.

Taken together with the shift-work literature, these results indicate that some subjective and objective measures (here alertness and body temperature) show a strong dependency on the endogenous regulation systems, as they display a circadian trend in different shiftwork conditions, like in controlled laboratory conditions, despite minor variations in this trend by external factors. This then implies that working during the circadian trough requires an additional effort as operators' functional state is at its lowest level. Most interestingly, other measures (here subjective tension and heart rate), which are known to display a circadian trend in controlled laboratory conditions, are much more influenced by external factors, so that the trends may vary considerably between situations (or jobsituations). Andorre and Queinnec (1998) reported a significant increase on the first shifthour for real-job performance (pages checked on a computer-screen) in operators controlling a chemical production process. In both studies the atypical trend was interpreted as indicating enhanced cognitive demands following shift take-over in jobsituations concerned with supervisory control of a dynamic process. Thus, in shift-work conditions, some psychological and physiological measures, and in particular those that are and other stress-sensitive, would be largely influenced by environmental factors, including meal-timing, task demands, time-pressure and so on, which may mask the otherwise circadian trend of these measures (Averty et al., 2004; Brookings et al., 1996; Khaleque, 1984; Rose et al., 1982).

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 43

> < 4 hours > 6 hours

< 4 hours > 6 hours

Therefore, 15 volunteer ATCs were asked to rate Thayer's (1989) checklist on 01:00, 07:00, 13:00, 19:00 by indicating whether they were on shift for four hours at most or for six hours at least (table 2). Statistical analyses revealed significant time-of-day and time-on-duty effects for both measures. ATCs rated alertness at a lower level at 01:00 and 07:00 than at 13:00 and 19:00 and tension at 07:00 compared to 01:00, 13:00 and 19:00. However, while selfrated tension was higher following long on-shift time, the opposite pattern was observed for self-rated alertness. In other words, when controllers started day-duty they experienced high alertness and low tension, whereas they reported decreased alertness and higher

**Figure 2.** Mean (+/-SE) subjective alertness (upper pannel) and tension (lower pannel) in ATCs as a

13:00 19:00

13:00 19:00

These data favour the interpretation that lower alertness reported by 12-h workers compared to 8-h workers on the early afternoon resulted from the fact that the former were on the second half of duty while the latter started their afternoon-duty (Tucker et al., 1998). Reduced day time alertness was observed on shifts starting late in the morning, though controllers had probably sufficient sleep on the night prior the shift. This raises the

subjective tension after several hours on duty.

**Index of Subjective Alertness**

1 1,2 1,4 1,6 1,8 2 2,2 2,4 2,6 2,8 3

5

5,5

6

6,5

**Index of Subjective Tension**

7

7,5

8

function of recording time (Mélan et al., 2007).

**Figure 1.** Upper panel: Mean (+/-S.E.) alertness level on 3 occasions within each of three shifts (1h following shift-start, middle, and 1h before shift-end). Lower panel: Mean (+/-S.E.) sublingual temperature on the same recordings (Cariou et al., 2008).

#### **3.3. Circadian variations of subjective measures in ATCs**

In light of the inconsistencies of the findings in the literature concerning other subjective measures than alertness, we investigated whether ATCs displayed typical circadian trends for subjective measures and whether on-shift time would modulate these measures. In this job-situation, traffic density variations across the 24h-day determine, partly at least, shift schedules that include in particular overlapping shifts and variable shift-duration.


**Table 2.** Time on duty of controllers on each of six shifts at the time each recording was performed (Mélan et al., 2007).

Therefore, 15 volunteer ATCs were asked to rate Thayer's (1989) checklist on 01:00, 07:00, 13:00, 19:00 by indicating whether they were on shift for four hours at most or for six hours at least (table 2). Statistical analyses revealed significant time-of-day and time-on-duty effects for both measures. ATCs rated alertness at a lower level at 01:00 and 07:00 than at 13:00 and 19:00 and tension at 07:00 compared to 01:00, 13:00 and 19:00. However, while selfrated tension was higher following long on-shift time, the opposite pattern was observed for self-rated alertness. In other words, when controllers started day-duty they experienced high alertness and low tension, whereas they reported decreased alertness and higher subjective tension after several hours on duty.

42 Advances in Air Navigation Services

**Figure 1.** Upper panel: Mean (+/-S.E.) alertness level on 3 occasions within each of three shifts (1h following shift-start, middle, and 1h before shift-end). Lower panel: Mean (+/-S.E.) sublingual

In light of the inconsistencies of the findings in the literature concerning other subjective measures than alertness, we investigated whether ATCs displayed typical circadian trends for subjective measures and whether on-shift time would modulate these measures. In this job-situation, traffic density variations across the 24h-day determine, partly at least, shift

Shift 01:00 07:00 13:00 19:00

09:00 - 20:00 4h00 10h00 11:00 - 20:00 2h00 8h00 15:30 - 23:00 3h30

Time of recording

schedules that include in particular overlapping shifts and variable shift-duration.

06:30 -14:00 0h30 6h30 07:00 - 17:30 0h00 6h00

**Table 2.** Time on duty of controllers on each of six shifts at the time each recording was performed

temperature on the same recordings (Cariou et al., 2008).

0,5

8h

10h30

13h

15h

18h

21h

23h

2h30

6h

1,5

2,5

3,5

20:00 - 07:00 5h00

(Mélan et al., 2007).

**3.3. Circadian variations of subjective measures in ATCs** 

**Figure 2.** Mean (+/-SE) subjective alertness (upper pannel) and tension (lower pannel) in ATCs as a function of recording time (Mélan et al., 2007).

These data favour the interpretation that lower alertness reported by 12-h workers compared to 8-h workers on the early afternoon resulted from the fact that the former were on the second half of duty while the latter started their afternoon-duty (Tucker et al., 1998). Reduced day time alertness was observed on shifts starting late in the morning, though controllers had probably sufficient sleep on the night prior the shift. This raises the

#### 44 Advances in Air Navigation Services

possibility that the chronic sleep deficit observed in shift-workers may, partly, account for decreased day-time alertness.

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 45

> 6 items auditory 9 items auditory 6 items visual 9 items visual

6 items auditory 6 items visual 9 items auditory 9 items visual

while varying item modality (auditory or visual) during encoding and restitution, list-length

**Figure 3.** Mean number of correct responses in a probe recognition task (upper panel) and in a free recall task (lower panel) on each of four recordings according to the number of items presented (6- vs. 9-

01:00 07:00 13:00 19:00

01:00 07:00 13:00 19:00

Overall, ATCs' performance was lowest on 07:00 compared to 13:00 and 19:00, in particular for visual items, for the longer lists and in the free recall task (Fig.3). In consequence, when the tasks involved more demanding processing performance was decreased particularly when alertness was low. Interestingly, significant main effects occurred for time-of-day, listlength and modality with the free recall procedure, while the three variables interacted when restitution was based on a recognition procedure. In that case, performance dropped in the early morning only for visual 9-item lists, i.e. in the more demanding task conditions. In the literature, differential time-of-day effects on participants' recall performance in recognition and free recall tasks were accounted for by similar processing differences (Folkard, 1979; Folkard, et al., 1976; Folkard & Monk, 1980; Lorenzetti & Natale, 1996; Oakhill & Davies, 1989). Accordingly, the finding of an overall effect on free recall but not on recognition performance may indicate that external factors (time of day) more readily

item lists) and presentation modality (auditory vs. visual; Mélan, et al., 2007).

(6- and 9-item lists) and restitution processes (recognition and free recall).

2,4 2,6 2,8 3 3,2 3,4 3,6 3,8

0,4 0,45 0,5 0,55 0,6 0,65 0,7 0,75 0,8 0,85 0,9

**Correct probability**

**Correct responses**

Further, both measures were negatively correlated, indicating that the lower ATCs quoted alertness, the higher they quoted tension. The data thus extend the findings of a circadian variation of subjective measures in ATCs and they favour the interpretation that in stressrelated job-situations enhanced tension may compensate for decreased alertness (Thayer, 1989), thereby enabling the maintenance of safety. Further investigations will however be necessary to establish firmly whether ATCs' tension-ratings indicate a direct influence by environmental factors (i.e. hours at work, heavy traffic), or whether the observed variations are merely the consequence of their functional state, as indicated by the negative correlation between tension and alertness. Thus, air traffic control activities differ indeed between day- and night-work, as high traffic on day-time involves sustained periods of high task requirements and attentional demands, whereas low traffic during the night would favor boredom proneness and sleepiness (Costa, 1999; Lille & Cheliout, 1982; Luna et al., 1997).

In recent years, most studies exploring the impact of shift work on health and well-being have reported troubles in psychological and social well-being, performance efficiency and increased stress levels (Costa, 2003). Some of these issues will be highlighted in the next section by exploring the relationship between ATCs' physiological state and their performance efficiency during short-term recall of verbal material.

### **4. Complex interactions between task-related factors and ATCs' functional state**

The shift-work literature shows that psychological measures differ not only according to time of day but also according to task characteristics, with high performance for tasks requiring rather automatic processing such as short-term memory task when alertness is low (in the morning), and high performance in more demanding tasks, relying for instance on long-term memory processing when alertness is high (in the evening). Laboratory studies reported indeed higher immediate recall in short-term memory tasks in the morning and enhanced delayed recall from long-term memory in the evening (Folkard, 1979; Folkard et al., 1976; Folkard & Monk, 1980; Monk & Embrey, 1981). Further, recall performance was reported to be higher in the morning when a recognition procedure was used, and in the afternoon when participants had to recall the text with a more demanding free recall procedure (Lorenzetti & Natale, 1996; Oakhill & Davies, 1989). Free recall would be cognitively more demanding as it involves an active search of memory traces in the absence of any cues, while item recognition would rely on less demanding item familiarity (Brébion et al., 2005; Mandler, 1980; Prince et al., 2005).

In light of the findings of the literature, it was important to investigate whether and to what extent the above-reported modality effect may be sensitive to alertness variations, but also to the cognitive effort required to remember the verbal material (Mélan, et al., 2007). ATCs' recall performance was recorded on different times of the day (01:00, 07:00, 13:00 and 19:00) while varying item modality (auditory or visual) during encoding and restitution, list-length (6- and 9-item lists) and restitution processes (recognition and free recall).

44 Advances in Air Navigation Services

1982; Luna et al., 1997).

**functional state** 

decreased day-time alertness.

possibility that the chronic sleep deficit observed in shift-workers may, partly, account for

Further, both measures were negatively correlated, indicating that the lower ATCs quoted alertness, the higher they quoted tension. The data thus extend the findings of a circadian variation of subjective measures in ATCs and they favour the interpretation that in stressrelated job-situations enhanced tension may compensate for decreased alertness (Thayer, 1989), thereby enabling the maintenance of safety. Further investigations will however be necessary to establish firmly whether ATCs' tension-ratings indicate a direct influence by environmental factors (i.e. hours at work, heavy traffic), or whether the observed variations are merely the consequence of their functional state, as indicated by the negative correlation between tension and alertness. Thus, air traffic control activities differ indeed between day- and night-work, as high traffic on day-time involves sustained periods of high task requirements and attentional demands, whereas low traffic during the night would favor boredom proneness and sleepiness (Costa, 1999; Lille & Cheliout,

In recent years, most studies exploring the impact of shift work on health and well-being have reported troubles in psychological and social well-being, performance efficiency and increased stress levels (Costa, 2003). Some of these issues will be highlighted in the next section by exploring the relationship between ATCs' physiological state and their

The shift-work literature shows that psychological measures differ not only according to time of day but also according to task characteristics, with high performance for tasks requiring rather automatic processing such as short-term memory task when alertness is low (in the morning), and high performance in more demanding tasks, relying for instance on long-term memory processing when alertness is high (in the evening). Laboratory studies reported indeed higher immediate recall in short-term memory tasks in the morning and enhanced delayed recall from long-term memory in the evening (Folkard, 1979; Folkard et al., 1976; Folkard & Monk, 1980; Monk & Embrey, 1981). Further, recall performance was reported to be higher in the morning when a recognition procedure was used, and in the afternoon when participants had to recall the text with a more demanding free recall procedure (Lorenzetti & Natale, 1996; Oakhill & Davies, 1989). Free recall would be cognitively more demanding as it involves an active search of memory traces in the absence of any cues, while item recognition would rely on less demanding item familiarity (Brébion

In light of the findings of the literature, it was important to investigate whether and to what extent the above-reported modality effect may be sensitive to alertness variations, but also to the cognitive effort required to remember the verbal material (Mélan, et al., 2007). ATCs' recall performance was recorded on different times of the day (01:00, 07:00, 13:00 and 19:00)

performance efficiency during short-term recall of verbal material.

et al., 2005; Mandler, 1980; Prince et al., 2005).

**4. Complex interactions between task-related factors and ATCs'** 

**Figure 3.** Mean number of correct responses in a probe recognition task (upper panel) and in a free recall task (lower panel) on each of four recordings according to the number of items presented (6- vs. 9 item lists) and presentation modality (auditory vs. visual; Mélan, et al., 2007).

Overall, ATCs' performance was lowest on 07:00 compared to 13:00 and 19:00, in particular for visual items, for the longer lists and in the free recall task (Fig.3). In consequence, when the tasks involved more demanding processing performance was decreased particularly when alertness was low. Interestingly, significant main effects occurred for time-of-day, listlength and modality with the free recall procedure, while the three variables interacted when restitution was based on a recognition procedure. In that case, performance dropped in the early morning only for visual 9-item lists, i.e. in the more demanding task conditions.

In the literature, differential time-of-day effects on participants' recall performance in recognition and free recall tasks were accounted for by similar processing differences (Folkard, 1979; Folkard, et al., 1976; Folkard & Monk, 1980; Lorenzetti & Natale, 1996; Oakhill & Davies, 1989). Accordingly, the finding of an overall effect on free recall but not on recognition performance may indicate that external factors (time of day) more readily

#### 46 Advances in Air Navigation Services

impacted on ATCs' task performance when deeper processing was required to solve a task. As indicated above, item recognition requires less in-depth processing of the to-beremembered material than free recall (Brébion et al., 2005; Mandler, 1980; Prince et al., 2005). Thus, task-dependent factors (modality, list-length and recall procedure in the present case) even further impact on ATCs' performance when their functional state is low.

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 47

events would have led to decreased performance in the morning, while no such effect was observed in the afternoon, when alertness and thus available mental resources were high.

**Figure 4.** Graphic representation of putative relationships between cognitive load factors and cognitive

This model, summarized in figure 4, further indicates that the different cognitive load measures used in the study, i.e. subjective measures (self-rated effort), behavioural measures (correct responses) and psychophysiological measures (heart rate variability) display a differential sensitivity to the three kinds of load factors investigated. More especially, heart rate variability increased with germane load, operationalized by high self-reported

load categories (Galy et al., 2012).

#### **5. Discussion**

The main finding of the present contribution is that ATCs' performance depends in a complex manner from the task to be performed, but also from job organisation (shift schedules, shift-duration) and from physiological aspects (alertness, sensory and cognitive processing). Even though it is difficult to generalize between different job-situations, given in particular differences between activities, to-be-performed tasks and shift-scheduling, the data confirm findings reported in other safety-related job-situations for security agents in a nuclear power plant and for operators controlling satellites (Costa, 1999; Cariou et al., 2008; Galy et al., 2008). Working during the night causes a mismatch between the endogenous circadian timing system and the environmental synchronizers (the light/dark cycle in particular), with consequent disturbances of the normal circadian rhythms of psychophysiological functions, beginning with the sleep/wake rhythm, and thus operators' alertness. In addition to the disruptive effects of shift work on performance efficiency, its impact on health and well-being are now well-documented (Costa, 2003; Siegrist, 2010). In this respect, some international directives have recently stressed the need for the careful organization of shift and night work and the protection of shift workers' health.

Elsewhere, the present findings favour the idea of a more general model to account for the complex interactions reported so far. In this regard, we recently extended Sweller's cognitive load theory elaborated in the educational field (Sweller, 1988; 1994), to a real-job situation (Galy et al., 2012).

#### **5.1. Towards an integrated model of mental load during ATC?**

Sweller defined three categories of cognitive load in order to embrace the complexities of a given task, the conditions in which it is performed, and subject-related variables. "Intrinsic cognitive load" refers to the load induced by the material to be processed, such as task difficulty that is defined in particular by the number of items to be processed, and by item interactivity (Ayres, 2006; Kalyuga et al., 2003; Sweller & Chandler, 1994). "Extraneous mental workload" refers to the load induced by external factors, including work situation, work organization, time pressure, and noise (Sweller, 1994). Finally, "germane mental workload" corresponds to the load induced by conscious application of strategies to solve tasks more easily (Schnotz & Kürschner, 2007). We showed an additive effect of intrinsic load (task difficulty) and extrinsic load (time pressure) factors, and that this effect was only observed in the morning. In other words, when participants performed a difficult task under high time pressure in the morning, when alertness and thus available mental resources were low, they probably had to use specific strategies generating germane load. This sequence of events would have led to decreased performance in the morning, while no such effect was observed in the afternoon, when alertness and thus available mental resources were high.

46 Advances in Air Navigation Services

**5. Discussion** 

situation (Galy et al., 2012).

impacted on ATCs' task performance when deeper processing was required to solve a task. As indicated above, item recognition requires less in-depth processing of the to-beremembered material than free recall (Brébion et al., 2005; Mandler, 1980; Prince et al., 2005). Thus, task-dependent factors (modality, list-length and recall procedure in the present case)

The main finding of the present contribution is that ATCs' performance depends in a complex manner from the task to be performed, but also from job organisation (shift schedules, shift-duration) and from physiological aspects (alertness, sensory and cognitive processing). Even though it is difficult to generalize between different job-situations, given in particular differences between activities, to-be-performed tasks and shift-scheduling, the data confirm findings reported in other safety-related job-situations for security agents in a nuclear power plant and for operators controlling satellites (Costa, 1999; Cariou et al., 2008; Galy et al., 2008). Working during the night causes a mismatch between the endogenous circadian timing system and the environmental synchronizers (the light/dark cycle in particular), with consequent disturbances of the normal circadian rhythms of psychophysiological functions, beginning with the sleep/wake rhythm, and thus operators' alertness. In addition to the disruptive effects of shift work on performance efficiency, its impact on health and well-being are now well-documented (Costa, 2003; Siegrist, 2010). In this respect, some international directives have recently stressed the need for the careful

even further impact on ATCs' performance when their functional state is low.

organization of shift and night work and the protection of shift workers' health.

**5.1. Towards an integrated model of mental load during ATC?** 

Elsewhere, the present findings favour the idea of a more general model to account for the complex interactions reported so far. In this regard, we recently extended Sweller's cognitive load theory elaborated in the educational field (Sweller, 1988; 1994), to a real-job

Sweller defined three categories of cognitive load in order to embrace the complexities of a given task, the conditions in which it is performed, and subject-related variables. "Intrinsic cognitive load" refers to the load induced by the material to be processed, such as task difficulty that is defined in particular by the number of items to be processed, and by item interactivity (Ayres, 2006; Kalyuga et al., 2003; Sweller & Chandler, 1994). "Extraneous mental workload" refers to the load induced by external factors, including work situation, work organization, time pressure, and noise (Sweller, 1994). Finally, "germane mental workload" corresponds to the load induced by conscious application of strategies to solve tasks more easily (Schnotz & Kürschner, 2007). We showed an additive effect of intrinsic load (task difficulty) and extrinsic load (time pressure) factors, and that this effect was only observed in the morning. In other words, when participants performed a difficult task under high time pressure in the morning, when alertness and thus available mental resources were low, they probably had to use specific strategies generating germane load. This sequence of

**Figure 4.** Graphic representation of putative relationships between cognitive load factors and cognitive load categories (Galy et al., 2012).

This model, summarized in figure 4, further indicates that the different cognitive load measures used in the study, i.e. subjective measures (self-rated effort), behavioural measures (correct responses) and psychophysiological measures (heart rate variability) display a differential sensitivity to the three kinds of load factors investigated. More especially, heart rate variability increased with germane load, operationalized by high self-reported alertness, whereas participants' self-rated mental effort was sensitive to both extrinsic and intrinsic load and task performance was determined both by alertness and by an interaction between task difficulty and time pressure. The latter results stress the difficulties encountered in order to identify a reliable measure of cognitive load on one hand (Backs & Seljos, 1994; Brookings et al., 1996; Carroll et al., 1986), and suggest that it may be hazardous to generalize the proposed model to other situations, on the other hand.

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 49

Frankish (2008), all models of speech perception incorporate some form of auditory short term memory, because speech comprehension requires the integration of information from successive elements. The findings reported here may thus be useful for various control operations and in particular during conflict resolution, when controllers' memory span may be more readily taxed. As shown by several studies, short-term recall of navigational messages decreases when message length increases beyond three commands (Wickens & Hollands, 2000; Barshi & Healy, 2002; Schneider et al., 2004). Taken together with several other studies, the present results may also be useful for ATC selection and training, as they stress the importance of using tests that manipulate, in addition to the more traditional quantitative aspects of memory, more qualitative parameters, such as presentation modality

or information type and aircraft status (Gronlund et al., 2005; Means et al., 1988).

or integration of graphical representations.

have been reviewed recently (Cowan, 2011).

term memory for problem resolution.

ATC, like several other safety-related job-activities rely preferentially on visual interfaces, most probably because graphical representations enhance the understanding of complex principles or spatial relations, for instance flight direction, speed and other information that an operator needs to synthetize in order to solve conflicting situations for instance. Within the auditory stream, successive items are strongly associated; in contrast, in the visual modality, it is simultaneously presented items that are strongly associated (Penney, 1989). Speech, with its linear progression through time does not bear these properties. In addition, sound and speech generate considerable noise when compared to silent reading of messages

However, for precisely the same reason, i.e. the disturbance induced by auditory information, this modality is typically used for alarm devices, as one cannot avoid an auditory signal. Indeed, the physical nature of sounds, i.e. waves that cross space in all directions, ensures that this kind of information automatically reaches and stimulates the auditory receptors and is transferred thereafter to the auditory cortex. There is thus a pretty good chance for heard messages or signals to be processed, even without paying attention to one's environment or when a subject's motivation or alertness are decreased for some reason. Furthermore, in addition to a specific activation of the auditory system, sounds also induce a non-specific activation of the nervous system, both the autonomous nervous system as evidenced for instance by increased heart-rate, and the central nervous system, as indicated by enhanced arousal. On contrary, written messages may be processed by the visual system only once a subject has oriented his/her gaze on the specific location where the message has occurred. Thus, unless the operator turns his/her attention to a particular location on a radar screen, he/she will not be able to appreciate whether for instance some airplane may be in a difficult position. Research concerning participant's focus of attention that would reflect conscious awareness and its relation to visual working memory tasks

In light of these considerations it appears that supervisory control greatly benefits from a well-weighted combination of visual representations of complex multi-dimensional data and auditory presentation of essential information which have to be maintained in a short-

With these limitations in mind, it may nevertheless be challenging to test the mental load model in ATCs, given that the findings reported in the previous sections indicate how work organization (shift schedules, shift-duration) affected alertness (germane mental load), and that when ATCs' alertness was low, their mnemonic performance was lower in the demanding task conditions (intrinsic mental load). Further, as conflicts between aircrafts require ATCs to make the right decisions under high time pressure and to give ground-toair instructions in a limited time, it seems indeed worthwhile to include this extrinsic load factor to a more integrative approach of mental load generated during ATC activities.

#### **5.2. Experimental designs to explore ATC activities**

Experimental paradigms like the ones used in the studies reported in this contribution may be regarded as simplified models of real-life activities. As such, the results reported may be of interest for ATC activities involving similar cognitive processes than those explored in the experimental paradigms. This is in particular the case for task-related factors, as the influence of such factors may only be demonstrated in controlled study conditions. In this regard, experimental approaches are complementary to *in situ* observations, which point to the relevant aspects to be explored more systematically by using experimental designs.

Further, the observations reported here were performed while ATCs were in their habitual work environment and work conditions. In contrast, in a number of field studies subjective data across the 24h-day are collected retrospectively, i.e. participants are asked to rate these measures on a single session by remembering what they perceived for instance during the night or during the morning shift. In contrast, subjective and performance measures reported in the present contribution were collected in real-time fashion, in order to gain some insight into ATCs' cognitive abilities in real-job conditions. This is particularly important when investigating job-activities performed on a continuous 24h-day, like ATC. The data reported here clearly show that shift-work features (time-of-day and time-on-shift) are indeed critical factors that impact not only on operators' functional state, but also on a number of psychological measures. ATCs' information processing ability is crucial for the safety of the air traffic management system as well as for the sector capacity of a given complexity in a particular time driving the overall system performances.

The present findings may be relevant to ensure productivity and/or safety in job-situations involving supervisory control, and in particular ATC, all relying on processing visual and auditory information from various sources on control panels and interfaces. The present findings suggest that short sequences of auditory information would less readily tax controller's processing capacities than longer sequences and/or visual material. As stated by Frankish (2008), all models of speech perception incorporate some form of auditory short term memory, because speech comprehension requires the integration of information from successive elements. The findings reported here may thus be useful for various control operations and in particular during conflict resolution, when controllers' memory span may be more readily taxed. As shown by several studies, short-term recall of navigational messages decreases when message length increases beyond three commands (Wickens & Hollands, 2000; Barshi & Healy, 2002; Schneider et al., 2004). Taken together with several other studies, the present results may also be useful for ATC selection and training, as they stress the importance of using tests that manipulate, in addition to the more traditional quantitative aspects of memory, more qualitative parameters, such as presentation modality or information type and aircraft status (Gronlund et al., 2005; Means et al., 1988).

48 Advances in Air Navigation Services

alertness, whereas participants' self-rated mental effort was sensitive to both extrinsic and intrinsic load and task performance was determined both by alertness and by an interaction between task difficulty and time pressure. The latter results stress the difficulties encountered in order to identify a reliable measure of cognitive load on one hand (Backs & Seljos, 1994; Brookings et al., 1996; Carroll et al., 1986), and suggest that it may be hazardous

With these limitations in mind, it may nevertheless be challenging to test the mental load model in ATCs, given that the findings reported in the previous sections indicate how work organization (shift schedules, shift-duration) affected alertness (germane mental load), and that when ATCs' alertness was low, their mnemonic performance was lower in the demanding task conditions (intrinsic mental load). Further, as conflicts between aircrafts require ATCs to make the right decisions under high time pressure and to give ground-toair instructions in a limited time, it seems indeed worthwhile to include this extrinsic load

factor to a more integrative approach of mental load generated during ATC activities.

Experimental paradigms like the ones used in the studies reported in this contribution may be regarded as simplified models of real-life activities. As such, the results reported may be of interest for ATC activities involving similar cognitive processes than those explored in the experimental paradigms. This is in particular the case for task-related factors, as the influence of such factors may only be demonstrated in controlled study conditions. In this regard, experimental approaches are complementary to *in situ* observations, which point to the relevant aspects to be explored more systematically by using experimental designs.

Further, the observations reported here were performed while ATCs were in their habitual work environment and work conditions. In contrast, in a number of field studies subjective data across the 24h-day are collected retrospectively, i.e. participants are asked to rate these measures on a single session by remembering what they perceived for instance during the night or during the morning shift. In contrast, subjective and performance measures reported in the present contribution were collected in real-time fashion, in order to gain some insight into ATCs' cognitive abilities in real-job conditions. This is particularly important when investigating job-activities performed on a continuous 24h-day, like ATC. The data reported here clearly show that shift-work features (time-of-day and time-on-shift) are indeed critical factors that impact not only on operators' functional state, but also on a number of psychological measures. ATCs' information processing ability is crucial for the safety of the air traffic management system as well as for the sector capacity of a given

The present findings may be relevant to ensure productivity and/or safety in job-situations involving supervisory control, and in particular ATC, all relying on processing visual and auditory information from various sources on control panels and interfaces. The present findings suggest that short sequences of auditory information would less readily tax controller's processing capacities than longer sequences and/or visual material. As stated by

complexity in a particular time driving the overall system performances.

to generalize the proposed model to other situations, on the other hand.

**5.2. Experimental designs to explore ATC activities** 

ATC, like several other safety-related job-activities rely preferentially on visual interfaces, most probably because graphical representations enhance the understanding of complex principles or spatial relations, for instance flight direction, speed and other information that an operator needs to synthetize in order to solve conflicting situations for instance. Within the auditory stream, successive items are strongly associated; in contrast, in the visual modality, it is simultaneously presented items that are strongly associated (Penney, 1989). Speech, with its linear progression through time does not bear these properties. In addition, sound and speech generate considerable noise when compared to silent reading of messages or integration of graphical representations.

However, for precisely the same reason, i.e. the disturbance induced by auditory information, this modality is typically used for alarm devices, as one cannot avoid an auditory signal. Indeed, the physical nature of sounds, i.e. waves that cross space in all directions, ensures that this kind of information automatically reaches and stimulates the auditory receptors and is transferred thereafter to the auditory cortex. There is thus a pretty good chance for heard messages or signals to be processed, even without paying attention to one's environment or when a subject's motivation or alertness are decreased for some reason. Furthermore, in addition to a specific activation of the auditory system, sounds also induce a non-specific activation of the nervous system, both the autonomous nervous system as evidenced for instance by increased heart-rate, and the central nervous system, as indicated by enhanced arousal. On contrary, written messages may be processed by the visual system only once a subject has oriented his/her gaze on the specific location where the message has occurred. Thus, unless the operator turns his/her attention to a particular location on a radar screen, he/she will not be able to appreciate whether for instance some airplane may be in a difficult position. Research concerning participant's focus of attention that would reflect conscious awareness and its relation to visual working memory tasks have been reviewed recently (Cowan, 2011).

In light of these considerations it appears that supervisory control greatly benefits from a well-weighted combination of visual representations of complex multi-dimensional data and auditory presentation of essential information which have to be maintained in a shortterm memory for problem resolution.

#### **Author details**

Claudine Mélan *Shift-work and Cognition Laboratory, Toulouse University, France* 

Edith Galy

*Research Centre in the Psychology of Cognition, Language, and Emotion, Aix-Marseille University, France* 

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 51

Brébion, G., David, A.S., Bressan, R.A. & Pilowski, L.S. (2005). Word frequency effects on free recall and recognition in patients with schizophrenia. *Journal of Psychiatry Research*,

Conrad, R. & Hull, A.J. (1968) Input modality and the serial position curve in short-term

Cariou, M.; Galy, E. & Mélan, C. (2008). Differential 24-h variations of alertness and subjective tension in process controllers: investigation of a relationship with body

Carroll, L.M.; Jalbert, A., Penney, A.M., Neath, I., Surprenant, A.M. & Tehan, G. (2010). Evidence for proactive interference in the focus of attention of working memory.

Costa, G. (1996). The impact of shift and nightwork on health. *Applied Ergonomics*, Vol.27,

Costa, G. (1999). Fatigue and biological rhythms. In: *Handbook of aviation human factors*. D.J.

Costa, G. (2003). Shift work and occupational medicine: An overview. *Occupational Medicine*,

Cowan, N. (1984). On short and long auditory stores. *Psychological Bulletin*, Vol.96, pp. 341-

Cowan, N. (2011).The focus of attention as observed in visual working memory tasks:

Cowan, N.; Saults, J.S., Elliott, E.M. & Moreno, M.V. (2002). Deconfounding serial recall.

Cowan, N.; Saults, J.S. & Brown, G. D. (2004). On the auditory modality superiority effect in serial recall: separating input and output factors. *Journal of Memory and Language*, Vol.46,

Crowder, R.G. & Morton, J. (1969). Precategorical acoustic storage (PAS). *Perception and* 

Frankish, C. (1985). Modality-specific grouping effects in short-term memory. *Journal of* 

Frankish, C. (2008). Precategorical acoustic storage and the perception of speech. *Journal of* 

Della Rocco, P.S. & Nesthus, T.E. (2005). Shift-work and air traffic control: Transitioning research results to the workforce. In: *Human factors impacts in air traffic management*. B.

Dinges, L.S.; Pack, F., Williams, W., Gillen, K.A., Powell, J.W., Ott, G.E., Aptowicz, C. & Pack, A.I. (1997). Cumulative sleepiness, mood disturbance, and psychomotor vigilance performance decrements during a week of sleep restricted to 4-5 hours per night. *Sleep*,

Kirwan, M. Rogers, D. Schäfer, (Eds), 243-278, Aldershot, UK: Ashgate

Making sense of competing claims. *Neuropsychologia*, Vol.49, pp. 1401–1406

*Journal of Memory and Language*, Vol.46, pp. 153-177

temperature and heart rate. *Chronobiology International*, Vol.25, No.4, pp. 97-609 Carroll, D., Turner J.R. & Hellawell, J.C. (1986). Heart rate and oxygen consumption during active psychological challenge: the effects of level of difficulty. Psychophysiology.

memory. *Psychonomic Science*, Vol.10, no.4), pp.135-136.

*Canadian Journal of Experimental Psychology*, Vol.64, pp. 208–214

Garland, J.A. Wise & V.D. Hopkin, (Eds.), 235-255, London: Erlbaum

Vol.39, pp. 215-222

23:174-181

no.1, pp. 9-16

Vol.53, pp. 83-88

pp. 153-177

*Psychophysics*, Vol.5, pp. 365-373

Vol.20, No.4, pp. 267-277

*Memory and Language*, Vol.24, pp. 200-209

*Memory and Language*, Vol.58, pp. 815–836

370

#### **6. References**


Brébion, G., David, A.S., Bressan, R.A. & Pilowski, L.S. (2005). Word frequency effects on free recall and recognition in patients with schizophrenia. *Journal of Psychiatry Research*, Vol.39, pp. 215-222

50 Advances in Air Navigation Services

*Shift-work and Cognition Laboratory, Toulouse University, France* 

*Behavior*, Vol.90, No.2-3, pp. 209-18

Environmental Medicine, Vol.75, pp. 333-341

*in Cognitive Science*, Vol.4, No.11, pp. 417–423

*Motivation*, G.A. Bower, (Ed.), 48–79, Academic Press

the modality effect in free recall. *Cognition*, Vol.77, 59-65

navigation task. *Memory and Cognition*, Vol.30, pp. 1189-1203

*Psychophysiology*, Vol.16, pp. 57-69

*and performance*, T. Monk, (Ed.), 129-152, John Wiley & Sons ltd,

alertness. *Aviation Space and Environmental Medicine,* Vol.75, pp.75-83

individual. *International Journal of Neuroscience*, Vol.52, pp. 29-37

within problems. *Learning and Instruction*, Vol.16, pp. 389–400

Baddeley, A. D. (1986). *Working memory*. Oxford: Oxford University Press, UK

*Research Centre in the Psychology of Cognition, Language, and Emotion, Aix-Marseille University,* 

Akerstedt, T. (1991). Sleepiness at work: Effect of irregular work hours. In: *Sleep, sleepiness* 

Akerstedt, T. (2007). Altered sleep/wake patterns and mental performance. *Physiology and* 

Akerstedt T, Folkard S. & Fortin C. (2004). Predictions from the three-process model of

Akerstedt, T. & Gillberg, M. (1990). Subjective and objective sleepiness in the active

Andorre, V. & Quéinnec, Y. (1998). Changes in supervisory activity of a continuous process during night and day shifts. *International Journal of Industrial Ergonomics*, Vol.21, pp. 179-

Averty, P., Collet, C., Dittmar, A., Athenes, S. & Vernet-Maury E. (2004). Mental workload in air traffic control: an index constructed from field tests. Aviation Space and

Ayres, P. (2006). Using subjective measures to detect variations of intrinsic cognitive load

Backs, R.W. & Seljos, K.A. (1994). Metabolic and cardio-respiratory of mental effort: the effects of level of difficulty in a working memory task. *International Journal of* 

Baddeley, A. D. (2000). The episodic buffer: a new component of working memory? *Trends* 

Baddeley, A. D. & Hitch, G. (1974). Working memory. In: *The Psychology of Learning and* 

Brookings, J.B, Wilson, G.F. & Swain, C.R. (1996). Psychophysiological responses to changes in workload during simulated air traffic control. *Biological Psychology*, Vol.42, pp. 361-

Beaman, C.P. & Morton, J. (2000). The separate but related origins of the recency effect and

Barshi, I. & Healy, A. F. (2002). The effects of mental representation on performance in a

**Author details** 

Claudine Mélan

**6. References** 

186

377

Edith Galy

*France* 


Folkard, S. (1979). Times of day and level of processing. *Memory and Cognition*, Vol.7, PP. 247-252.

Recall Performance in Air Traffic Controllers Across the 24-hr Day: Influence of Alertness and Task Demands on Recall Strategies 53

Luna, T.D., French. J. & Mitcha, J.L. (1997). A study of USAF air traffic controller shiftwork: sleep, fatigue, activity, and mood analyses. *Aviation Space and Environmental Medicine*,

Madigan, S. A. (1971). Modality and recall order interactions in short-term memory for serial

Means, B.; Mumaw, R.J., Roth, C., Schlager, M.S., McWilliams, E., Gagne, E. et al. (1988). *ATC training analysis study: Design of the next-generation ATC training system*. report N°. FAA/OPM 342-036, Washington, DC: Department of Transportation/federal Aviation

Mélan, C.; Galy, E. & Cariou, M. (2007). Mnemonic Processing in Air Traffic Controllers (ATCs): Effects of Task Parameters and Work Organization. *International Journal of* 

Monk, T.H.; Fookson, J.E., Moline, M.L. & Pollak, C.P. (1985). Diurnal variation in mood and performance in a time isolated environment. *Chronobiology International*, Vol.2, pp. 185-

Oakhill, J. & Davies, A.M. (1989). The effects of time of day and subjects' test expectations on recall and recognition of prose materials. *Acta Psychologica*, Vol.72, pp. 145-157 Oberauer, K. & Vockenberg, K. (2009). Updating of working memory: Lingering bindings.

Owens, D. S.; Macdonald, I., Tucker, P., Sytnik, N., Totterdell, P., Minors, D., et al. (2000). Diurnal variations in the mood and performance of highly practiced young women living under strictly controlled conditions. *British Journal of Psychology,* Vol.91, pp. 41-60 Penney, C. G. (1989). Modality effects and structure of short-term verbal memory. *Memory* 

Prizmic Z.; Vidadek S., Radosevic-Vidadek B. & Kaliterna L. (1995). Shiftwork tolerance and

Rose, R.M.; Jenkins, C.D., Hurst, M., Herd, J.A. & Hall RP. (1982). Endocrine activity in air traffic controllers at work. II. Biological, psychological and work correlates.

Rosa R. (1995). Extended workshifts and excessive fatigue. Journal of Sleep Research, Vol.4,

Rummer, R. & Schweppe, J. (2005). Evidence for a modality effect in sentence retention.

Prince, S.E., Daselaar, S.M. & Cabeza, R. (2005). Neural of relational memory: successful encoding and retrieval of semantic and perceptual associations. *Journal of Neuroscience*,

Schneider, V.I.; Healy, A.F. & Barshi, I. (2004). Effects of instruction modality and read-back on accuracy in flowing navigation commands. *Journal of Experimental Psychology:* 

Schnotz, W. & Kürschner, C. (2007). A reconsideration of cognitive load theory. *Educational* 

Siegrist, J. (2010). Effort-reward imbalance at work and cardiovascular diseases. *International* 

*Journal of Occupational Medicine and Environmental Health*, Vol.23, pp. 279-285

*The Quarterly Journal of Experimental Psychology*, Vol.62, pp. 967-987

24-h variations in moods. *Work Stress*, Vol. 9, pp. 327-334

*Psychonomic Bulletin & Review*, Vol.12, No.6, pp.1094-1099

order. *Journal of Experimental Psychology*, Vol.87, pp. 294-296

*Aviation Psychology*, Vol.17, No.4, pp. 391-409

*and Cognition*, Vol.17, No.4, pp. 398-422

*Psychoneuroendocrinology,* Vol.7, pp. 113-123

Vol.68, pp. 18-23

Administration, USA

193

pp. 51-56

Vol.25, pp. 1203-1210

*Applied*, Vol.10, pp. 245-257

*Psychology Review*, Vol.19, pp. 469–508


Luna, T.D., French. J. & Mitcha, J.L. (1997). A study of USAF air traffic controller shiftwork: sleep, fatigue, activity, and mood analyses. *Aviation Space and Environmental Medicine*, Vol.68, pp. 18-23

52 Advances in Air Navigation Services

*Ergonomics*, Vol.19, pp. 479-488.

*Psychology*, Vol.71, pp. 295-307.

26, Boston: Birkhaüser

Vol.12, No1, pp. 3-15

*Psychology*, Vol.15, pp. 269-293

*Cognition*, Vol.15, pp. 846-858

*Memory*, Vol.15, No.7, pp. 693-700

*Human Ergology*, Vol.13, pp. 23-29

sleep and alertness. *Sleep,* Vol. 20, pp. 215-223

*Educational Psychologist*, Vol.38, pp. 23–31

pp. 269-275

*(London)*, Vol.53, PP. 95-101.

Ergonomics, Vol.51, No.9, pp. 1338-51

*Aviation Psychology,* Vol.20, No.3, pp. 295-307.

247-252.

Folkard, S. (1979). Times of day and level of processing. *Memory and Cognition*, Vol.7, PP.

Folkard, S.; Knauth, P., Monk, T.H. & Rutenfranz, J. (1976). The effect of memory load on the circadian variation in performance efficiency under a rapidly rotating shift system.

Folkard, S. & Monk, T. H. (1980). Circadian rhythms in human memory. *British Journal of* 

Folkard, S. & Akerstedt, T. (1992). A three-process model of the regulation of alertnesssleepiness. In: *Sleep, arousal, and performance,* R.J. Broughton & R.D. Ogilvie, (Eds.), 11-

Folkard, S. & Tucker, P. (2003). Shift work, safety and productivity. *Occupational Medicine* 

Galy, E.; Cariou, M. & Mélan, C. (2012). What is the relationship between mental workload factors and cognitive load types?, *International Journal of Psychophysiology*, Vol.83, No.3,

Galy, E.; Mélan, C. & Cariou, M. (2008). Investigation of task performance variations according to task requirements and alertness across the 24-h day in shift-workers.

Galy, E.; Mélan, C. & Cariou, M. (2010). Investigation of ATCs' response strategies in a free recall task: what makes auditory recall superior to visual recall? *International Journal of* 

Glenberg, A. M., & Swanson, N. G. (1986). A temporal distinctiveness theory of recency and modality effects. *Journal of Experimental Psychology: Learning, Memory, and Cognition*,

Gronlund, S.D.; Dougherty, M.R.P., Durso, F.T., Canning, J.M. & Mills, S. H. (2005). Planning in air traffic control: Impact of problem type. *International Journal of Aviation* 

Gronlund, S.D.; Ohrt, D.D., Dougherty, M.R.P., Perry, J.L. & Manning, C.A. (1998). Role of memory in air traffic control. *Journal of Experimental Psychology: Learning, Memory, &* 

Harvey, A.J. & Beaman, C.P. Input and output modality effects in immediate serial recall.

Kecklund, G., Akerstedt, T. & Lowden, A. (1997). Morning work: effects of early rising on

Kalyuga, S., Ayres, P., Chandler, P. & Sweller, J. (2003). The expertise reversal effect.

Khaleque, A. (1984). Circadian rhythms in heart rate of shift and day workers. *Journal of* 

Lille, F. & Cheliout, F. (1982). Variations in diurnal and nocturnal waking state in air traffic

Lorenzetti, R. & Natale, V. (1996). Time of day and processing strategies in narrative

controllers. *European Journal of Applied Physiology*, Vol.49, pp. 319-328

comprehension. *British Journal of Psychology*, Vol.87, pp. 209-221


Sweller, J. (1988). Cognitive load during problem solving: effects on learning. *Cognitive Science*, Vol.12, pp. 257–285

**Chapter 4** 

© 2012 Diana et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2012 Diana et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**Predicting Block Time:** 

Additional information is available at the end of the chapter

1 The source is http://www.merriam-webster.com/dictionary/predict.

NEXTOR\_TDI\_Report\_Final\_October\_2010.pdf.

Tony Diana

http://dx.doi.org/10.577248622

well as risk measurement.

**1. Introduction** 

**An Application of Quantile Regression** 

According to Merriam Webster1, to predict is 'to declare or indicate in advance; foretell on the basis of observation, experience, or scientific reason.' The advent of sophisticated mathematical and statistical techniques has taken 'divination' out of prediction. In the late 19th century, the work of Francis Galton in the areas of regression analysis, correlation and the normal distribution has been instrumental in helping analysts investigate the relationship between dependent and independent variables and, as a result, to be able to improve forecast. More recently, economists such Robert Engle and Clive Granger have made significant contributions to the study of time series that have widespread applications nowadays in economics and especially finance, such as price and interest rate volatility, as

Aviation is another industry that faces risk and uncertainty and has greatly benefited by advances in mathematical, statistical and operations research techniques. A flight is an event that can be scheduled up to six months ahead of its execution. However, despite the best preparation, flight performance is subject to many factors beyond human control such as weather, equipment failure, labor actions, security threats, etc. As a main contributor to the economy and global trade, government regulators, airlines and airport authorities have a vested interest in ensuring that the aviation system supports unimpeded movements of goods and people from their origin to their final destination. According to the *Total Delay Impact Study*2 by a group of Nextor researchers, "the total cost of all US air transportation delays in 2007 was \$32.9 billion. The \$8.3 billion airline component consists of increased

2 Ball, M. et al., 2010. *Total delay impact study, a comprehensive assessment of the costs and impacts of flight delay in the United States*, Nextor, vii. The report is available at the following website: http://its.berkeley.edu/sites/default/files/


## **Predicting Block Time: An Application of Quantile Regression**

Tony Diana

54 Advances in Air Navigation Services

Vol.24, pp. 49-54

*Science*, Vol.12, pp. 257–285

*Instruction*, Vol.12, pp. 185–233

*Learning and Instruction,* Vol.4, pp. 295–312

ed.). Upper Saddle River, NJ: Prentice Hall, USA

Sweller, J. (1988). Cognitive load during problem solving: effects on learning. *Cognitive* 

Sweller, J. (1994). Cognitive load theory, learning difficulty and instructional design.

Sweller, J. & Chandler, P. (1994). Why some material is difficult to learn. *Cognition and* 

Thayer, R.E. (1989). *The biopsychology of mood and arousal*. New-York: Oxford University Press Tucker, P., Lombardi, D., Smith L. & Folkard, S. (2006). The Impact of Rest Breaks on Temporal Trends in Injury Risk. *Chronobiology International,* Vol. 23, pp. 1423–1434 Tucker P, Smith L, Macdonald I, Folkard S. (1998). Shift length as a determinant of retrospective on-shift alertness. *Scandinavian Journal of Work and Environmental Health*,

Wright, K.P., Hull, J.T. & Czeisler, C.A. (2002). Relationship between alertness, performance, and body temperature in humans. *American Journal of Physiology*, Vol. 253, pp. 1370-1377 Wickens, C.D. & Hollands, J.G. (2000). *Engineering psychology and human performance* (3rd Additional information is available at the end of the chapter

http://dx.doi.org/10.577248622

#### **1. Introduction**

According to Merriam Webster1, to predict is 'to declare or indicate in advance; foretell on the basis of observation, experience, or scientific reason.' The advent of sophisticated mathematical and statistical techniques has taken 'divination' out of prediction. In the late 19th century, the work of Francis Galton in the areas of regression analysis, correlation and the normal distribution has been instrumental in helping analysts investigate the relationship between dependent and independent variables and, as a result, to be able to improve forecast. More recently, economists such Robert Engle and Clive Granger have made significant contributions to the study of time series that have widespread applications nowadays in economics and especially finance, such as price and interest rate volatility, as well as risk measurement.

Aviation is another industry that faces risk and uncertainty and has greatly benefited by advances in mathematical, statistical and operations research techniques. A flight is an event that can be scheduled up to six months ahead of its execution. However, despite the best preparation, flight performance is subject to many factors beyond human control such as weather, equipment failure, labor actions, security threats, etc. As a main contributor to the economy and global trade, government regulators, airlines and airport authorities have a vested interest in ensuring that the aviation system supports unimpeded movements of goods and people from their origin to their final destination. According to the *Total Delay Impact Study*2 by a group of Nextor researchers, "the total cost of all US air transportation delays in 2007 was \$32.9 billion. The \$8.3 billion airline component consists of increased

<sup>2</sup> Ball, M. et al., 2010. *Total delay impact study, a comprehensive assessment of the costs and impacts of flight delay in the United States*, Nextor, vii. The report is available at the following website: http://its.berkeley.edu/sites/default/files/ NEXTOR\_TDI\_Report\_Final\_October\_2010.pdf.

<sup>1</sup> The source is http://www.merriam-webster.com/dictionary/predict.

<sup>© 2012</sup> Diana et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Diana et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

expenses for crew, fuel, and maintenance, among others. The \$16.7 billion passenger component is based on the passenger time lost due to schedule buffer, delayed flights, flight cancellations, and missed connections. The \$3.9 billion cost from lost demand is an estimate of the welfare loss incurred by passengers who avoid air travel as the result of delays".

Predicting Block Time: An Application of Quantile Regression 57

regression is more robust to outliers than the traditional OLS regression because the latter

This is of importance to aviation practitioners and, especially, airline schedulers who have often resorted to schedule padding in order to make up for ground and en route delays. This research presents a different perspective on the study of predictability with the intent to

To assess the impact of selected operational covariates at different locations of the

To derive more predictable block times based on the impact of operational covariates at

To test a model without any assumption about the distribution of errors and

After a brief background, the discussion will proceed with the methodology, an explanation

A focus group including communication navigation surveillance and air traffic management representatives4 defined predictability as "a measure of delay variance against a performance dependability target. As the variance of expected delay increases, it becomes a

According to Donohue et al. (2001:398), "predictability focuses on the variation in the ATM [Air Traffic Management] system as experienced by the user. Predictability includes both variability in flight times and arrival rates". In this article, the study of predictability is extended beyond wheels-off (takeoff) and wheels-off (landing) times to include any flight operations between gate-out and gate-in times such as taxi-out and taxi-in movements. This

For Vossen et al. (2011:388), "flexibility can be defined as the amount of operational latitude granted to the carriers in meeting their individual objectives (e.g. on-time arrival, network preservation, profit) when disruptions occur. […] The notion of predictability is closely related, and can be defined as the reduction of uncertainties in the implementation of ATFM [Air Traffic Flow Management] initiatives". Although airlines have to face many events in the course of a flight that cannot thoroughly be anticipated and planned for, "ATFM initiatives should provide the user with time to react, and the provider's intent should be

Predictability is sometimes associated with the concept of robust airline scheduling. The latter is the outcome of four sequential tasks as schedule generation, fleet assignment,

4 Report of the Air Traffic Services Performance Focus Group and Communication Navigation Surveillance, February 1999. *Airline Metric Concepts for Evaluating Air Traffic Service Performance.* The website is http://www.boeing.com/

very serious concern for airlines when developing and operating their schedules".

does not focus on the conditional mean.

distribution of block time.

of the outcomes and some final comments.

approach takes into account passenger experience.

communicated as clearly and as far in advance as possible".

commercial/caft/cwg/ats\_perf/ATSP\_Feb1\_Final.pdf.

various quantiles.

**2. Background** 

help aviation analysts achieve the following objectives:

homoscedasticity (constant variance of the residuals).

Predictability is all the more difficult to achieve as airlines often face three types of delay. First, delays can be induced: The air traffic control authority can initiate a ground delay program in case of adverse weather conditions or heavy traffic volume on the ground or enroute. Second, delays can be propagated: In a sequence of legs operated by the same tailnumbered aircraft, a flight may accumulate delays that cannot be recovered by the end of the itinerary. Finally, delays can be stochastic because they are the results of random events such as equipment breakdown or extreme weather events.

Predictability represents a key performance area in the aviation industry for several reasons.


Recently, much discussion has revolved around the validity of using airlines' schedules as a measure of on-time performance and the variance of block delay as an indicator of predictability. Both airlines' limited control over the three types of delay and airport congestion make it difficult to build robust schedules and to use schedule as a reference for on-time performance. In fact, schedule padding may skew actual airline performance assessment, hence the need for an alternative methodology.

This article proposes a methodology to determine the predictability of block time based on the case study of the Seattle-Oakland city pair. The predictable block time is located at the percentile where the sign and magnitude of the pseudo coefficient of determination is the highest, while all the covariates are significant at a given confidence level. Ordinary-leastsquare (OLS) regression models enable analysts to evaluate the percentage of variation in actual block time explained by changes in selected operational variables. However, quantile

<sup>3</sup> Henk J. Hof, Development of a Performance Framework in support of the Operational Concept, ICAO Mid Region Global ATM Operational Concept Training Seminar, Cairo, Egypt, November 28–December 1, 2005, p. 36.

regression is more robust to outliers than the traditional OLS regression because the latter does not focus on the conditional mean.

This is of importance to aviation practitioners and, especially, airline schedulers who have often resorted to schedule padding in order to make up for ground and en route delays. This research presents a different perspective on the study of predictability with the intent to help aviation analysts achieve the following objectives:


After a brief background, the discussion will proceed with the methodology, an explanation of the outcomes and some final comments.

#### **2. Background**

56 Advances in Air Navigation Services

reasons.

expenses for crew, fuel, and maintenance, among others. The \$16.7 billion passenger component is based on the passenger time lost due to schedule buffer, delayed flights, flight cancellations, and missed connections. The \$3.9 billion cost from lost demand is an estimate of the welfare loss incurred by passengers who avoid air travel as the result of delays".

Predictability is all the more difficult to achieve as airlines often face three types of delay. First, delays can be induced: The air traffic control authority can initiate a ground delay program in case of adverse weather conditions or heavy traffic volume on the ground or enroute. Second, delays can be propagated: In a sequence of legs operated by the same tailnumbered aircraft, a flight may accumulate delays that cannot be recovered by the end of the itinerary. Finally, delays can be stochastic because they are the results of random events

Predictability represents a key performance area in the aviation industry for several

 For the International Civil Aviation Organization (ICAO), predictability refers to the "ability of the airspace users and ATM service providers to provide consistent and

 One of the goals of the U.S. Next Generation of Air Transportation System (NextGen) is to foster the transition from an air traffic control to more of an air traffic managed system where pilots have more flexibility to select their routes, utilize performancebased navigation (PBN) with the help of satellites and make decisions based on

 According to Rapajic (2009:51), "cutting five minutes of average of 50 per cent of schedules thanks to higher predictability would be worth some €1,000 million per annum, through savings or better use of airlines and airport resources." Unpredictability imposes considerable costs on airlines in the forms of lost revenues,

Recently, much discussion has revolved around the validity of using airlines' schedules as a measure of on-time performance and the variance of block delay as an indicator of predictability. Both airlines' limited control over the three types of delay and airport congestion make it difficult to build robust schedules and to use schedule as a reference for on-time performance. In fact, schedule padding may skew actual airline performance

This article proposes a methodology to determine the predictability of block time based on the case study of the Seattle-Oakland city pair. The predictable block time is located at the percentile where the sign and magnitude of the pseudo coefficient of determination is the highest, while all the covariates are significant at a given confidence level. Ordinary-leastsquare (OLS) regression models enable analysts to evaluate the percentage of variation in actual block time explained by changes in selected operational variables. However, quantile

3 Henk J. Hof, Development of a Performance Framework in support of the Operational Concept, ICAO Mid Region

Global ATM Operational Concept Training Seminar, Cairo, Egypt, November 28–December 1, 2005, p. 36.

customer dissatisfaction and potential loss of market share.

assessment, hence the need for an alternative methodology.

such as equipment breakdown or extreme weather events.

dependable levels of performance."3

automated information-sharing.

A focus group including communication navigation surveillance and air traffic management representatives4 defined predictability as "a measure of delay variance against a performance dependability target. As the variance of expected delay increases, it becomes a very serious concern for airlines when developing and operating their schedules".

According to Donohue et al. (2001:398), "predictability focuses on the variation in the ATM [Air Traffic Management] system as experienced by the user. Predictability includes both variability in flight times and arrival rates". In this article, the study of predictability is extended beyond wheels-off (takeoff) and wheels-off (landing) times to include any flight operations between gate-out and gate-in times such as taxi-out and taxi-in movements. This approach takes into account passenger experience.

For Vossen et al. (2011:388), "flexibility can be defined as the amount of operational latitude granted to the carriers in meeting their individual objectives (e.g. on-time arrival, network preservation, profit) when disruptions occur. […] The notion of predictability is closely related, and can be defined as the reduction of uncertainties in the implementation of ATFM [Air Traffic Flow Management] initiatives". Although airlines have to face many events in the course of a flight that cannot thoroughly be anticipated and planned for, "ATFM initiatives should provide the user with time to react, and the provider's intent should be communicated as clearly and as far in advance as possible".

Predictability is sometimes associated with the concept of robust airline scheduling. The latter is the outcome of four sequential tasks as schedule generation, fleet assignment,

<sup>4</sup> Report of the Air Traffic Services Performance Focus Group and Communication Navigation Surveillance, February 1999. *Airline Metric Concepts for Evaluating Air Traffic Service Performance.* The website is http://www.boeing.com/ commercial/caft/cwg/ats\_perf/ATSP\_Feb1\_Final.pdf.

#### 58 Advances in Air Navigation Services

aircraft routing and crew pairing/rostering (Wu 2010; Abdelghany and Abdelghany 2009). Fleet assignment models (FAM) are often used to determine how demand for air travel is met by available fleet (see Abara 1989 and Hane et al. 1995). Moreover, the fleet assignment models present two challenges: complexity and size of the problem that the FAM can handle.

Predicting Block Time: An Application of Quantile Regression 59

skewness coefficients6 are 0.11, -0.44, 0.37, and 0.19 respectively for summer 2011, 2010, 2004 and 2000. A negative skew indicates that the left tail is longer. While the standard deviation is appropriate to measure the spread of a symmetric distribution, interquartile ranges are

Secondly, summer is part of the high travel season when demand is usually at its peak. This, in turn, is likely to increase airport congestion and subsequently impact block time. Finally, the years were selected to account for (1) pre- and post-September 11, 2001 traffic, (2) lower traffic demand resulting from the 2008-2009 economic recession, and (3) the introduction of

The key performance indicators of flight performance are summarized in Table 1. Although the number of flights increased between 2000 and 2011 and the average minutes of expected departure clearance times (EDCT) were higher in 2011 than in 2000, the percentage of ontime gate departures and arrivals and other key delay indicators such as taxi-out delay (a measure of ground congestion) improved in 2011. It is interesting to point out that the percentage of flights in IMC did not change significantly at OAK among the four selected summers. IMC operations were, however, much higher in 2010 and 2011 than in 2000 at

The sample does not include a variable that measures performance-based navigation. The available surveillance data such as Traffic Flow Management System (TFMS) do not capture

6 The skewness coefficient is computed as γ = E[(x – μ)3/σ] = μ3/σ3 where μ3 is the third moment about the mean μ and

7 The Green Skies over Seattle program includes initiatives such as reduced track mileage to minimum possible distance to protect the environment, optimized profile descent, reduction or elimination of low altitude radar

SEA, which explains the existence of average minutes of EDCT in 2010 and 2011.

more indicative of spread changes in skewed distributions (see Figure 1).

**Figure 1.** Box Plots of Actual Block Times (SEA-OAK)

the Green Skies over Seattle7 after 2010.

σ is the standard deviation and E is the expectation operator.

vectoring, as well as required navigational performance.

Rapajic (2009) identified network structure and fleet composition as sources of flight irregularities. Wu (2010) provided an excellent exposition of issues related to delay management, operating process optimization, and schedule disruption management. Wu explained that "airline schedule planning is deeply rooted in economic principles and market forces, some of which are imposed and constrained by the operating environment of the [airline] industry" (2010:11). He presented a schedule optimization model to improve the robustness of airline scheduling. However, such a model does not consider how selective operational variables are likely to impact scheduling.

Morrisset and Odoni (2011) compared runway system capacity, air traffic delay, scheduling practices, and flight schedule reliability at thirty-four major airports in Europe and the United States from 2007 to 2008. The authors explained that European airports limit air traffic delay through slot control. The other difference is that declared capacity (therefore, the number of available slots) is based mainly on operations under instrument meteorological conditions (IMC). By not placing any restrictions on the number of operations, schedule reliability in the United States depends more on weather conditions than at European airports.

#### **3. Methodology**

#### **3.1. The sample and the assumptions**

The sample includes daily data for the month of June to August in 2000, 2004, 2010 and 2011 for the Seattle/Tacoma International (SEA)-Oakland International (OAK) city pair. The summer season is usually characterized by low ceiling and visibility that determine instrument meteorological conditions5 and weather events such as thunderstorms—all likely to skew the distribution of block times.

Illustration 1 compares the boxplots of actual block times in minutes for the four summers under investigation. The boxplot shows the spread of the distribution, the selected quantile values, the position of the mean and median block times, and the presence of outliers that make it important to consider a regression model at different quantiles. The boxplots reveal an increase in the actual block times between summer 2004 and 2011. Summer 2010 features the largest range as well as the lowest block times at the 5th percentile among the four samples (Illustration 1). It is also characterized by the highest proportion of operations in instrument meteorological conditions compared with the other three samples (Table 1). The

<sup>5</sup> The minimum ceiling and visibility at SEA are respectively 4,000 feet and 3 nautical miles. At OAK, they are 2,500 feet and 8 nautical miles.

skewness coefficients6 are 0.11, -0.44, 0.37, and 0.19 respectively for summer 2011, 2010, 2004 and 2000. A negative skew indicates that the left tail is longer. While the standard deviation is appropriate to measure the spread of a symmetric distribution, interquartile ranges are more indicative of spread changes in skewed distributions (see Figure 1).

**Figure 1.** Box Plots of Actual Block Times (SEA-OAK)

58 Advances in Air Navigation Services

than at European airports.

**3.1. The sample and the assumptions** 

to skew the distribution of block times.

**3. Methodology** 

and 8 nautical miles.

operational variables are likely to impact scheduling.

handle.

aircraft routing and crew pairing/rostering (Wu 2010; Abdelghany and Abdelghany 2009). Fleet assignment models (FAM) are often used to determine how demand for air travel is met by available fleet (see Abara 1989 and Hane et al. 1995). Moreover, the fleet assignment models present two challenges: complexity and size of the problem that the FAM can

Rapajic (2009) identified network structure and fleet composition as sources of flight irregularities. Wu (2010) provided an excellent exposition of issues related to delay management, operating process optimization, and schedule disruption management. Wu explained that "airline schedule planning is deeply rooted in economic principles and market forces, some of which are imposed and constrained by the operating environment of the [airline] industry" (2010:11). He presented a schedule optimization model to improve the robustness of airline scheduling. However, such a model does not consider how selective

Morrisset and Odoni (2011) compared runway system capacity, air traffic delay, scheduling practices, and flight schedule reliability at thirty-four major airports in Europe and the United States from 2007 to 2008. The authors explained that European airports limit air traffic delay through slot control. The other difference is that declared capacity (therefore, the number of available slots) is based mainly on operations under instrument meteorological conditions (IMC). By not placing any restrictions on the number of operations, schedule reliability in the United States depends more on weather conditions

The sample includes daily data for the month of June to August in 2000, 2004, 2010 and 2011 for the Seattle/Tacoma International (SEA)-Oakland International (OAK) city pair. The summer season is usually characterized by low ceiling and visibility that determine instrument meteorological conditions5 and weather events such as thunderstorms—all likely

Illustration 1 compares the boxplots of actual block times in minutes for the four summers under investigation. The boxplot shows the spread of the distribution, the selected quantile values, the position of the mean and median block times, and the presence of outliers that make it important to consider a regression model at different quantiles. The boxplots reveal an increase in the actual block times between summer 2004 and 2011. Summer 2010 features the largest range as well as the lowest block times at the 5th percentile among the four samples (Illustration 1). It is also characterized by the highest proportion of operations in instrument meteorological conditions compared with the other three samples (Table 1). The

5 The minimum ceiling and visibility at SEA are respectively 4,000 feet and 3 nautical miles. At OAK, they are 2,500 feet

Secondly, summer is part of the high travel season when demand is usually at its peak. This, in turn, is likely to increase airport congestion and subsequently impact block time. Finally, the years were selected to account for (1) pre- and post-September 11, 2001 traffic, (2) lower traffic demand resulting from the 2008-2009 economic recession, and (3) the introduction of the Green Skies over Seattle7 after 2010.

The key performance indicators of flight performance are summarized in Table 1. Although the number of flights increased between 2000 and 2011 and the average minutes of expected departure clearance times (EDCT) were higher in 2011 than in 2000, the percentage of ontime gate departures and arrivals and other key delay indicators such as taxi-out delay (a measure of ground congestion) improved in 2011. It is interesting to point out that the percentage of flights in IMC did not change significantly at OAK among the four selected summers. IMC operations were, however, much higher in 2010 and 2011 than in 2000 at SEA, which explains the existence of average minutes of EDCT in 2010 and 2011.

The sample does not include a variable that measures performance-based navigation. The available surveillance data such as Traffic Flow Management System (TFMS) do not capture

<sup>6</sup> The skewness coefficient is computed as γ = E[(x – μ)3/σ] = μ3/σ3 where μ3 is the third moment about the mean μ and σ is the standard deviation and E is the expectation operator.

<sup>7</sup> The Green Skies over Seattle program includes initiatives such as reduced track mileage to minimum possible distance to protect the environment, optimized profile descent, reduction or elimination of low altitude radar vectoring, as well as required navigational performance.

#### 60 Advances in Air Navigation Services

whether a pilot had requested a required navigation performance procedure, whether air traffic control had granted the request, and whether the procedure had actually been implemented. Moreover, it is presently difficult to differentiate flown performance-based navigation procedures from instrument landing system (ILS) approaches in the case of flight track overlay.

Predicting Block Time: An Application of Quantile Regression 61

**Actual Block Time** (ACTBLKTM) is the dependent variable. It refers to the time in

 **Block Buffer** (BLKBUFFER) represents the difference between planned and optimal block time. The latter is the sum of unimpeded taxi-out times and filed estimated time enroute. Block buffer is the additional minutes included in planned block time in order to take into account potential induced, propagated and stochastic delays. It has also been defined as "the additional time built into the schedule specifically to absorb delay whilst the aircraft is on the ground and to allow recovery between the rotations of aircraft" (Cook, 2007:105). Donohue et al. (2001:113) explained that "to obtain their desired on-time performance, airlines will add padding into a schedule to reflect an amount above average block times to allow for delay and seasonally experienced

**Departure Delay** (DEPDEL) corresponds to difference between the actual and planned

**Arrival Delay** (ARRDEL) represents to the difference between the actual and planned

 **Airborne Delay** (AIRBNDEL) accounts for the total minutes of airborne delay. It is the difference between the actual airborne times (landing minus takeoff times) minus the

**Taxi-Out Time** (TXOUTTM) refers to the duration in minutes from gate departure to

Readers interested in quantile regression are referred to Hao and Naiman (2007), Koenker (2005), Koenker and Hallock (2001) and Koenker and Bassett (1998). Quantile regression provides several advantages compared with the ordinary-least-square (OLS) regression in assessing the influence of selected operational factors on the variations of block time at

 Quantile regression specifies the conditional quantile function and, therefore, a way to assess the probability of achieving a certain level of performance. It permits the analysis of the full conditional distributional properties of block time as opposed to ordinary-

 It defines functional relations between variables for all portions of a probability distribution. Quantile regression can improve the predictive relationship between block times and selected variables by focusing on quantiles instead of the mean. As Hao and Naiman (2007:4) pointed out, "While the linear regression model specifies the changes in the conditional mean of the dependent variable associated with a change in the covariates, the quantile regression model specifies changes in the conditional quantile." It determines the effect of explanatory variables on the central or non-central location,

 It is distribution-free, which allows the study of extreme quantiles. Outliers influence the length of the right tail and make average block time irrelevant as a standard for

least-square (OLS) regression models that focus on the mean.

scale, and shape of the distribution of block times.

minutes from actual gate departure to actual gate arrival.

gate departure time at the departure airport in a city pair.

gate arrival time at the arrival airport in a city pair.

variations in block times."

filed estimated time enroute.

various locations of its distribution:

**3.3. Quantile regression** 

wheels-off times (gate-out to wheels-off).


**Table 1.** Performance Metrics for the SEA-OAK City Pair

Secondly, the availability of Q-routes makes it possible for RNAV/RNP capable aircraft to reduce mileage, to minimize conflicts between routes and to maximize high-altitude airspace. Q-routes are available for use by RNAV/RNP capable aircraft between 18,000 feet MSL (mean sea level) and FL (flight level) 450 inclusive. They help minimize mileage and reduce conflicts between routes.

Thirdly, block time as a measure of gate-to-gate performance is sensitive to delays on the ground and en route. To account for this, airborne delay represents a surrogate for enroute congestion, while increases in taxi times imply surface movement congestion.

#### **3.2. Sources and definition of the variables**

The sources for the variables are ARINC8's Out-Off-On-In times and the U.S. Federal Aviation Administration's Traffic Flow Management System (TFMS). The directional city pair data originated from the 'Enroute' and 'Individual Flights' data marts of the Aviation System Performance Metrics (ASPM) data warehouse9.

The choice of variables reflects operational and statistical considerations. On the one hand, some model variables represent significant factors in airport congestion (taxi times) and enroute performance (airborne delays). On the other hand, the model with the highest values for the Akaike Information Criterion (AIC)10 and Bayesian Information Criterion (BIC)11 was selected in order to prevent overfitting and to reduce the number of covariates.

The dependent (response variable) and independent variables (covariates) are defined as follows:

<sup>8</sup> AIRINC stands for Aeronautical Radio, Inc. (http://www.arinc.com).

<sup>9</sup> The TFMS (formerly ETMS) and ARINC data as well as the ASPM delay metrics are available at http://aspm.faa.gov.

<sup>10</sup> The Akaike Information Criterion is defined as 2k – 2 ln(L) where k is the number of parameters and L the maximized value of the likelihood function for the estimated model.

<sup>11</sup> The Bayesian Information Criterion is -2 ln(L) + k.ln(n) where n is the number of observations.


#### **3.3. Quantile regression**

60 Advances in Air Navigation Services

**Table 1.** Performance Metrics for the SEA-OAK City Pair

**3.2. Sources and definition of the variables** 

System Performance Metrics (ASPM) data warehouse9.

8 AIRINC stands for Aeronautical Radio, Inc. (http://www.arinc.com).

maximized value of the likelihood function for the estimated model.

reduce conflicts between routes.

follows:

track overlay.

whether a pilot had requested a required navigation performance procedure, whether air traffic control had granted the request, and whether the procedure had actually been implemented. Moreover, it is presently difficult to differentiate flown performance-based navigation procedures from instrument landing system (ILS) approaches in the case of flight

Secondly, the availability of Q-routes makes it possible for RNAV/RNP capable aircraft to reduce mileage, to minimize conflicts between routes and to maximize high-altitude airspace. Q-routes are available for use by RNAV/RNP capable aircraft between 18,000 feet MSL (mean sea level) and FL (flight level) 450 inclusive. They help minimize mileage and

Thirdly, block time as a measure of gate-to-gate performance is sensitive to delays on the ground and en route. To account for this, airborne delay represents a surrogate for enroute

The sources for the variables are ARINC8's Out-Off-On-In times and the U.S. Federal Aviation Administration's Traffic Flow Management System (TFMS). The directional city pair data originated from the 'Enroute' and 'Individual Flights' data marts of the Aviation

The choice of variables reflects operational and statistical considerations. On the one hand, some model variables represent significant factors in airport congestion (taxi times) and enroute performance (airborne delays). On the other hand, the model with the highest values for the Akaike Information Criterion (AIC)10 and Bayesian Information Criterion (BIC)11 was selected in order to prevent overfitting and to reduce the number of covariates.

The dependent (response variable) and independent variables (covariates) are defined as

9 The TFMS (formerly ETMS) and ARINC data as well as the ASPM delay metrics are available at http://aspm.faa.gov. 10 The Akaike Information Criterion is defined as 2k – 2 ln(L) where k is the number of parameters and L the

11 The Bayesian Information Criterion is -2 ln(L) + k.ln(n) where n is the number of observations.

congestion, while increases in taxi times imply surface movement congestion.

Readers interested in quantile regression are referred to Hao and Naiman (2007), Koenker (2005), Koenker and Hallock (2001) and Koenker and Bassett (1998). Quantile regression provides several advantages compared with the ordinary-least-square (OLS) regression in assessing the influence of selected operational factors on the variations of block time at various locations of its distribution:


identifying the best-possible block time. A single rate of change characterized by the slope of the OLS regression line cannot be representative of the relationship between an independent variable or covariate and the entire distribution of block time, the response variable. In the quantile regression, the estimates represent the rates of change conditional on adjusting for the effects of the other model variables at a specified percentile. Therefore, the skewed distribution of block times calls for a more robust regression method that takes into account outliers or the lack of sufficient data at a particular percentile (especially at the extremes of the distribution) and generates different slopes for different quantiles.

Predicting Block Time: An Application of Quantile Regression 63

(1)

at a 95% confidence level in the 2011 and 2010 samples. However, since the intercept is

Among the dependent variables, gate arrival and departure delays are not significant at a 95% confidence level in the 2010 sample. This implies that airlines can make up for ground delays once en route or ground delays are likely to be more significant in a few extreme cases. The F statistics suggest that there is a zero percent chance that the dependent variable estimates are equal to zero. A value of the Durbin Watson statistic close to 2.00 suggests that there is little statistical evidence that the error terms are positively auto-correlated. The values of the coefficients of determination (R2) imply that the model covariates explain a

In the quantile regression models, the covariate estimates, as well as the key regression statistics at the 5th, 25th, median, 75th and 95th percentile, are summarized in the appendix 2 table. The 50th quantile estimates can be used to track location changes. According to Hao and Naiman (2007: 55), the 5th and 95th percentiles "can be used to assess how a covariate predicts the conditional off-central locations as well as shape shifts of the response." In the case of the 50th percentile in summer 2011, the quantile regression model for at τ (tau) = 0.50

> 0.50 BLKBUFFER SCHEDBLKTM DEPDEL ARRDEL AIRBNDEL TXOUTTM

 

Block Time 14.8091 0.8909 \* X 1.0038 \* X 0.3329 \* X 0.2936 \* X 1.0957 \* X 1.1606 \* X

In equation (1), 1.1606 represents the change in the median of block time between SEA and OAK corresponding to a one minute change in taxi-out time at SEA. Since the p value is zero, we reject the null hypothesis, at a 95 percent confidence level, that taxi-out times at SEA has no effect on the median block time between SEA and OAK in summer 2011. The pseudo coefficient of determination is a goodness-of-fit measure12. In the case of summer 2011, 80.21% of the variation in block time is explained by the model covariates at the 50th

No sample includes covariates that are significant at a 95% confidence level at all quantiles. Gate departure and arrival delays are significant only at the 95th percentile across the four samples. This means that departure and arrival delays are more likely to affect consistently block times in the upper percentiles—in case of severe airport congestion, for instance. Moreover, the magnitude of block buffers and gate departure delays have a negative impact on the conditional quantile of block time at all samples' selected percentiles. The size of the buffer and the time an aircraft will spend on the tarmac before take-off are conditions likely to affect block times. As a result, there is a need for analysts to decompose and to measure the different operations between gate-out and wheels-off times including gate departure, push-back, taxi-out and queuing times before wheels off. Airport Surface Detection

12 See Koenker and Machado 1999 for further explanations. According to Fitzenberger et al. (2010: 234), "the pseudo R2 equals one minus the sum of weighted deviations about estimated quantile over the sum of weighted deviations

necessary to provide more accurate predictions, it was left in the model.

high proportion of the variation in block times.

(50th percentile or median) is as follows:

percentile of block time (appendix 1).

around raw quantile".

The difference between OLS and quantile regression characteristics are summarized in the table below:


**Table 2.** The Assumptions of Linear and Quantile Regression

#### **4. Outcomes and implications**

Appendix 1 provides the estimates for the OLS models. The intercept that represents the predicted value of actual block time when the covariates are equal to zero is not significant at a 95% confidence level in the 2011 and 2010 samples. However, since the intercept is necessary to provide more accurate predictions, it was left in the model.

62 Advances in Air Navigation Services

table below:

different slopes for different quantiles.

 Estimates the mean of a response variable conditional on the values of the explanatory variables (specifies the

 Determines the rate of change in the mean of the response variable

 Provides a measure of the impact of explanatory variables on the central location of the distribution of the

 Does not account for full conditional distribution properties of the response

Assumption of constant variance in errors

**4. Outcomes and implications** 

**Table 2.** The Assumptions of Linear and Quantile Regression

conditional mean function)

response variable.

variable

identifying the best-possible block time. A single rate of change characterized by the slope of the OLS regression line cannot be representative of the relationship between an independent variable or covariate and the entire distribution of block time, the response variable. In the quantile regression, the estimates represent the rates of change conditional on adjusting for the effects of the other model variables at a specified percentile. Therefore, the skewed distribution of block times calls for a more robust regression method that takes into account outliers or the lack of sufficient data at a particular percentile (especially at the extremes of the distribution) and generates

The difference between OLS and quantile regression characteristics are summarized in the

Normal distribution (sensitive to outliers) Distribution-free (allows study of extreme

Normal distribution of errors No assumption about the distribution of

Appendix 1 provides the estimates for the OLS models. The intercept that represents the predicted value of actual block time when the covariates are equal to zero is not significant

(homoscedasticity) Does not assume homoscedasticity

Determines best fitting line for all data Different estimates for different

quantiles)

quantiles

errors

 Specifies the conditional quantile function (focus on quantiles). Defines functional relations between variables for all portions of a

 Determines the effect of explanatory variables on the central or non-central location, scale, and shape of the distribution of the response variable Permits the analysis of the full conditional distributional properties of

probability distribution

the dependent variable

Linear Regression Quantile Regression

Among the dependent variables, gate arrival and departure delays are not significant at a 95% confidence level in the 2010 sample. This implies that airlines can make up for ground delays once en route or ground delays are likely to be more significant in a few extreme cases. The F statistics suggest that there is a zero percent chance that the dependent variable estimates are equal to zero. A value of the Durbin Watson statistic close to 2.00 suggests that there is little statistical evidence that the error terms are positively auto-correlated. The values of the coefficients of determination (R2) imply that the model covariates explain a high proportion of the variation in block times.

In the quantile regression models, the covariate estimates, as well as the key regression statistics at the 5th, 25th, median, 75th and 95th percentile, are summarized in the appendix 2 table. The 50th quantile estimates can be used to track location changes. According to Hao and Naiman (2007: 55), the 5th and 95th percentiles "can be used to assess how a covariate predicts the conditional off-central locations as well as shape shifts of the response." In the case of the 50th percentile in summer 2011, the quantile regression model for at τ (tau) = 0.50 (50th percentile or median) is as follows:

$$\begin{aligned} \text{Block Time}\_{\text{r}=0.50} &= -14.8091 - 0.8909 \, ^\circ \text{X}\_{\text{BLKQUER}} + 1.0038 \, ^\circ \text{X}\_{\text{SCHDBLKTM}} - \\ 0.3329 \, ^\circ \text{X}\_{\text{DEDEL}} &+ 0.2936 \, ^\circ \text{X}\_{\text{ARDEL}} + 1.0957 \, ^\circ \text{X}\_{\text{AlBNDEL}} + 1.1606 \, ^\circ \text{X}\_{\text{TNOUTM}} + \varepsilon \end{aligned} \tag{1}$$

In equation (1), 1.1606 represents the change in the median of block time between SEA and OAK corresponding to a one minute change in taxi-out time at SEA. Since the p value is zero, we reject the null hypothesis, at a 95 percent confidence level, that taxi-out times at SEA has no effect on the median block time between SEA and OAK in summer 2011. The pseudo coefficient of determination is a goodness-of-fit measure12. In the case of summer 2011, 80.21% of the variation in block time is explained by the model covariates at the 50th percentile of block time (appendix 1).

No sample includes covariates that are significant at a 95% confidence level at all quantiles. Gate departure and arrival delays are significant only at the 95th percentile across the four samples. This means that departure and arrival delays are more likely to affect consistently block times in the upper percentiles—in case of severe airport congestion, for instance. Moreover, the magnitude of block buffers and gate departure delays have a negative impact on the conditional quantile of block time at all samples' selected percentiles. The size of the buffer and the time an aircraft will spend on the tarmac before take-off are conditions likely to affect block times. As a result, there is a need for analysts to decompose and to measure the different operations between gate-out and wheels-off times including gate departure, push-back, taxi-out and queuing times before wheels off. Airport Surface Detection

<sup>12</sup> See Koenker and Machado 1999 for further explanations. According to Fitzenberger et al. (2010: 234), "the pseudo R2 equals one minus the sum of weighted deviations about estimated quantile over the sum of weighted deviations around raw quantile".

#### 64 Advances in Air Navigation Services

Equipment, Model X or ASDE-X data should help do so as the system relying on a combination of surface movement radar and transponder multi-lateration sensors becomes more widespread.

Predicting Block Time: An Application of Quantile Regression 65

in ecology and biology than in the transportation industry, quantile regression is seldom

First, it enables aviation analysts to consider the impact of selected covariates on different locations of the distribution of block times. Secondly, the significance and the strength of the impact of selected covariates on block times make it possible to assess the probability that gateto-gate operations is likely to reach a specific duration. This is made possible by looking at the conditional quantile in the case of quantile regression as opposed to the conditional mean of the distribution of block times in the case of OLS models. Thirdly, quantile regression makes it easier to evaluate the scale and magnitude of change across specific percentiles over a sample. Finally, quantile regression can help analysts study the impact of covariates from different perspectives. For instance, in summer 2011, the data analysis suggests that 95% of the block time distribution will be below the quantile dependent variable value of 129.14 minutes as a result of the impact of the covariates' impact. Quantile regression enables the identification of more realistic threshold times based on quantiles and it allows airline practitioners to simulate

Predictability is a key performance area identified by the International Civil Aviation Organization. Moreover, it is a corner stone of the Next Generation of Air Transport System (NextGen) initiatives in the U.S. and the Single European Sky ATM Research program (SESAR) to ensure the transition from an air traffic controlled to a more air traffic managed environment. As air transportation regulators are under public pressure to crack down on tarmac and other types of delays, it has become imperative for airline schedulers to evaluate models that reflect the predictable influence of key operational variables on actual on-time performance. The complexity of the air traffic system, the inability for airline schedulers to fully anticipate both airport and en route congestion, and delays all make it more significant for aviation practitioners to assess the impact of some key operational variables at different locations of the distribution of block times that usually

The imbalance between air travel demand and airport capacity usually results in delays. As block times become more predictable, it is more possible for airline and airport operators to optimize airport capacity— especially at large congested airports. This is all the more significant in the U.S. where arrival and departure flows are not slot- constrained as in Europe. Block time predictability does not only affect how airports and airlines operate, but also the capability of air traffic control authorities to anticipate staff workload, as well as the ability of ground handlers

to minimize aircraft turn times by allocating resources where and when needed.

*Performance and Outreach, ANG-F1, SW, Washington DC, USA* 

15 Note: This article does not represent the opinion of the Federal Aviation Administration.

*Division Manager, NextGen Performance, Federal Aviation Administration, Office of NextGen* 

featured in econometric textbooks. Nevertheless, it presents several advantages.

and to evaluate various scenarios linked to changes in the models' covariates.

tends to be skewed due to outliers.

**Author details** 

Tony Diana15

Taking the example of summer 2011, 95 percent of the distribution of block times between SEA and OAK was below 129.14 minutes compared with a mean of 120.62 minutes (appendix 2). In other words, there is a 95 percent chance that actual block time will be lower than 129.14 minutes―based on a quantile regression model that explains 85.34 percent of the variation in block times. One benefit of quantile regression is that it facilitates the evaluation of scale and magnitude changes across samples and percentiles.

The quantile regression estimates in appendix 1 imply that block times increased in between summer 2000 and 2011 at all quantiles. In a comparison of summer 2000 with summer 2011, there had been an increase of 2.21 minutes in block times at the 95th percentile, for instance. The SEA-OAK city pair has been mainly operated by Southwest Airlines (SWA) and Alaska Airlines (ASA) with a predominant fleet of Boeing 737s. The total number of ASA arrivals and departures declined to 356 in summer 2011 from 693 in summer 2000―with 91 ASA ights operated by Horizon's Bombardier Q40013. Nevertheless, ASA operated larger capacity models such as the dash 400, 800 and 900 series, while SWA utilized a combination of dash 300, 500 and 700 models. The reason for the increase in block time may be attributed to airlines' operations policy to slow aircraft speed in order to save on fuel costs14. Weather conditions characterized by the percentage of operations in instrument meteorological conditions (IMC) did not vary substantially at OAK compared with SEA (see Table 1).

In appendix 3, the graphs illustrate the 95% confidence bands in the case of summer 2000. The estimates show a positive relationship between the quantile value and the estimated coefficients for scheduled block times, taxi out times and airborne delay, with a stronger effect in the upper tail. The effect of gate departure and arrival delays is not relatively constant, especially at the 50th percentile as implied by the wider bands around the 50th percentile value. These graphs are important for the analysts in identifying the quantiles where quantile value is likely to be close to the estimated coefficients and, therefore, to improve the accuracy of predicted block time.

#### **5. Final comments**

Based on the analysis of the SEA-OAK city pair case study, this research showed how quantile regression can help aviation practitioners develop more robust schedules. Originally proposed by Koenker and Bassett (1978), quantile regression is a rather novel approach to the analysis of airlines' on-time performance. Although it is more widely used

<sup>13</sup> The sources for schedules and aircraft mix are the Official Airline Guide (http://www.oag.com) and Innovata (http://www.innovata-llc.com).

<sup>14</sup> Associated Press. *Airlines slow down flights to save on fuel: JetBlue adds 2 minutes to each flight, saves \$13.6 million a year in jet fuel*, May 1, 2008. The article is available at the following website: http://www.msnbc.msn.com/id/24410809/ ns/business-us\_business/t/airlines-slow-down-flights-save-fuel/#.T01rmPES2Ag

in ecology and biology than in the transportation industry, quantile regression is seldom featured in econometric textbooks. Nevertheless, it presents several advantages.

First, it enables aviation analysts to consider the impact of selected covariates on different locations of the distribution of block times. Secondly, the significance and the strength of the impact of selected covariates on block times make it possible to assess the probability that gateto-gate operations is likely to reach a specific duration. This is made possible by looking at the conditional quantile in the case of quantile regression as opposed to the conditional mean of the distribution of block times in the case of OLS models. Thirdly, quantile regression makes it easier to evaluate the scale and magnitude of change across specific percentiles over a sample. Finally, quantile regression can help analysts study the impact of covariates from different perspectives. For instance, in summer 2011, the data analysis suggests that 95% of the block time distribution will be below the quantile dependent variable value of 129.14 minutes as a result of the impact of the covariates' impact. Quantile regression enables the identification of more realistic threshold times based on quantiles and it allows airline practitioners to simulate and to evaluate various scenarios linked to changes in the models' covariates.

Predictability is a key performance area identified by the International Civil Aviation Organization. Moreover, it is a corner stone of the Next Generation of Air Transport System (NextGen) initiatives in the U.S. and the Single European Sky ATM Research program (SESAR) to ensure the transition from an air traffic controlled to a more air traffic managed environment. As air transportation regulators are under public pressure to crack down on tarmac and other types of delays, it has become imperative for airline schedulers to evaluate models that reflect the predictable influence of key operational variables on actual on-time performance. The complexity of the air traffic system, the inability for airline schedulers to fully anticipate both airport and en route congestion, and delays all make it more significant for aviation practitioners to assess the impact of some key operational variables at different locations of the distribution of block times that usually tends to be skewed due to outliers.

The imbalance between air travel demand and airport capacity usually results in delays. As block times become more predictable, it is more possible for airline and airport operators to optimize airport capacity— especially at large congested airports. This is all the more significant in the U.S. where arrival and departure flows are not slot- constrained as in Europe. Block time predictability does not only affect how airports and airlines operate, but also the capability of air traffic control authorities to anticipate staff workload, as well as the ability of ground handlers to minimize aircraft turn times by allocating resources where and when needed.

#### **Author details**

64 Advances in Air Navigation Services

more widespread.

with SEA (see Table 1).

**5. Final comments** 

(http://www.innovata-llc.com).

improve the accuracy of predicted block time.

ns/business-us\_business/t/airlines-slow-down-flights-save-fuel/#.T01rmPES2Ag

Equipment, Model X or ASDE-X data should help do so as the system relying on a combination of surface movement radar and transponder multi-lateration sensors becomes

Taking the example of summer 2011, 95 percent of the distribution of block times between SEA and OAK was below 129.14 minutes compared with a mean of 120.62 minutes (appendix 2). In other words, there is a 95 percent chance that actual block time will be lower than 129.14 minutes―based on a quantile regression model that explains 85.34 percent of the variation in block times. One benefit of quantile regression is that it facilitates

The quantile regression estimates in appendix 1 imply that block times increased in between summer 2000 and 2011 at all quantiles. In a comparison of summer 2000 with summer 2011, there had been an increase of 2.21 minutes in block times at the 95th percentile, for instance. The SEA-OAK city pair has been mainly operated by Southwest Airlines (SWA) and Alaska Airlines (ASA) with a predominant fleet of Boeing 737s. The total number of ASA arrivals and departures declined to 356 in summer 2011 from 693 in summer 2000―with 91 ASA ights operated by Horizon's Bombardier Q40013. Nevertheless, ASA operated larger capacity models such as the dash 400, 800 and 900 series, while SWA utilized a combination of dash 300, 500 and 700 models. The reason for the increase in block time may be attributed to airlines' operations policy to slow aircraft speed in order to save on fuel costs14. Weather conditions characterized by the percentage of operations in instrument meteorological conditions (IMC) did not vary substantially at OAK compared

In appendix 3, the graphs illustrate the 95% confidence bands in the case of summer 2000. The estimates show a positive relationship between the quantile value and the estimated coefficients for scheduled block times, taxi out times and airborne delay, with a stronger effect in the upper tail. The effect of gate departure and arrival delays is not relatively constant, especially at the 50th percentile as implied by the wider bands around the 50th percentile value. These graphs are important for the analysts in identifying the quantiles where quantile value is likely to be close to the estimated coefficients and, therefore, to

Based on the analysis of the SEA-OAK city pair case study, this research showed how quantile regression can help aviation practitioners develop more robust schedules. Originally proposed by Koenker and Bassett (1978), quantile regression is a rather novel approach to the analysis of airlines' on-time performance. Although it is more widely used

13 The sources for schedules and aircraft mix are the Official Airline Guide (http://www.oag.com) and Innovata

14 Associated Press. *Airlines slow down flights to save on fuel: JetBlue adds 2 minutes to each flight, saves \$13.6 million a year in jet fuel*, May 1, 2008. The article is available at the following website: http://www.msnbc.msn.com/id/24410809/

the evaluation of scale and magnitude changes across samples and percentiles.

Tony Diana15 *Division Manager, NextGen Performance, Federal Aviation Administration, Office of NextGen Performance and Outreach, ANG-F1, SW, Washington DC, USA* 

<sup>15</sup> Note: This article does not represent the opinion of the Federal Aviation Administration.

### **Appendix**

#### **The ordinary-least square regression outputs**


Predicting Block Time: An Application of Quantile Regression 67

**The quantile regression outputs** 

#### **Summer 2000: Quantile process estimates (95% confidence level)**

#### **The quantile regression outputs**

66 Advances in Air Navigation Services

**The ordinary-least square regression outputs** 

**Summer 2000: Quantile process estimates (95% confidence level)** 


> -.2 .0 .2 .4 .6

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

ARRDEL

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

0.6 0.8 1.0 1.2 1.4

0.6 0.8 1.0 1.2 1.4 1.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

AIRBNDEL

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

SCHEDBLKTM

BLKBUFFER

Quantile Process Estimates (95% CI)

**Appendix** 


> -.6 -.5 -.4 -.3 -.2 -.1 .0 .1

0.5 0.6 0.7 0.8 0.9 1.0 1.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

DEPDEL

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

TXOUTTM

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Quantile

C


#### **6. References**

Abara, J., 1989. Applying integer linear programming to the fleet assignment problem. Interfaces 19(4), 20-28.

**Chapter 5** 

© 2012 Furuta et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2012 Furuta et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**Simulation of Team Cooperation Processes** 

Recent increase in air traffic demands makes the role of Air Traffic Control (ATC), which supports safety and efficiency of aviation, more important than ever. As aviation technologies have progressed, automation and computer supports are being introduced in cockpits, but ATC still heavily relies on human expertise of Air Traffic Control Officers (ATCOs). It is therefore necessary to understand ATC tasks from a viewpoint of ATCOs' cognitive behaviour in order to assess and improve task schemes and training programs

There are a few different phases in ATC, but this study exclusively focuses on en-route ATC. En-route ATC is performed by a team of ATCOs usually made up with a Radar Controller (RC) and a Coordination Controller (CC). RC monitors the radar screen, communicates with air pilots by radio, and gives instructions to them, while CC makes coordination with CCs in charge of other sectors, and supports RC. Team cooperation is therefore a key issue for good

We have been studying team cooperation processes in en-route ATC based on ethnographic field observation, and already proposed a cognitive model of team cooperation in en-route ATC as shown in Fig. 1 (Furuta et al., 2009; Soraji et al., 2010). In this model, establishment of Team Situation Awareness (TSA) on the air traffic is a key process for smooth cooperation. TSA here can be defined as a combination of individual situation awareness (Endsley, 1995) and mutual beliefs on it (Shu & Furuta, 2005). Once TSA has been established, individual tasks will be planned and executed almost implicitly, and TSA almost determines decision by ATCOs. The cognitive processes of a controller team after TSA acquisition are well described by a distributed version of the recognition-primed

control performance, and study on team cooperation processes is an important issue.

**in En-Route Air Traffic Control** 

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/48548

decision model (Klein, 1997).

**1. Introduction** 

for ATC.

Kazuo Furuta, Kouhei Ohno, Taro Kanno and Satoru Inoue


## **Simulation of Team Cooperation Processes in En-Route Air Traffic Control**

Kazuo Furuta, Kouhei Ohno, Taro Kanno and Satoru Inoue

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/48548

#### **1. Introduction**

68 Advances in Air Navigation Services

Interfaces 19(4), 20-28.

Abara, J., 1989. Applying integer linear programming to the fleet assignment problem.

Abdelghany, F. and Abdelghany, K., 2009. Modeling applications in the airline industry.

Cook, A., 2007. European air traffic management: principles, practices, and research.

Donohue, G.L., Zellweger, A., Rediess, H., and Pusch, C., 2001. Air transportation system engineering: progress in astronautics and aeronautics. American Institute of

Fitzenberger, B., Koenker, R., and Machado, J., 2010. Economic applications of quantile

Hane, C.A., Barnhart, C., Johnson, E.L., Marsten, R.E., Nemhauser, G.L., and Sigismondi, G., 1995. The fleet assignment problem: solving a large scale integer programming.

Hao, L. and Naiman, D. Q., 2007. Quantile Regression. Sage Publications: Thousand Oaks,

Koenker, R. and Machado, J., 1999. Goodness of fit and related inference processes for quantile regression, Journal of the American Statistical Association, 94(448), 1296-1310. Morissett, T. and Odoni, A., 2011. Capacity, delay, and schedule reliability at major airports in Europe and the United States. Transportation Research Record: Journal of the

Rapajic, J., 2009. Beyond airline disruptions. Ashgate Publishing Company: Burlington,

Vossen, T., Hoffman, R., and Mukherjee, A., 2011. Air traffic flow management in Quantitative problem solving methods in the airline industry: a modeling methodology

Wu, C.-L., 2010. Airline operations and delay management: insights from airline economics, networks and strategic schedule planning. Ashgate Publishing Company: Burlington,

handbook, Barnhart, C. and Smith, B., editors. Springer: New York.

Koenker, R., 2005. Quantile Regression. Cambridge, UK: Cambridge University Press. Koenker, R. and Hallock K., 2001. Quantile Regression. Journal of Economic Perspectives,

regression (Studies in empirical economics). Physica Verlag: Heidelberg.

Ashgate Publishing Company: Burlington, Vermont.

Ashgate Publishing Company: Burlington, Vermont.

Aeronautics and Astronautics: Danvers, Massachusetts.

Mathematical Programming, 70(2), 211-232.

Transportation Research Board, 2214, 85-93.

**6. References** 

CA.

15(4), 143-156.

Vermont.

Vermont.

Recent increase in air traffic demands makes the role of Air Traffic Control (ATC), which supports safety and efficiency of aviation, more important than ever. As aviation technologies have progressed, automation and computer supports are being introduced in cockpits, but ATC still heavily relies on human expertise of Air Traffic Control Officers (ATCOs). It is therefore necessary to understand ATC tasks from a viewpoint of ATCOs' cognitive behaviour in order to assess and improve task schemes and training programs for ATC.

There are a few different phases in ATC, but this study exclusively focuses on en-route ATC. En-route ATC is performed by a team of ATCOs usually made up with a Radar Controller (RC) and a Coordination Controller (CC). RC monitors the radar screen, communicates with air pilots by radio, and gives instructions to them, while CC makes coordination with CCs in charge of other sectors, and supports RC. Team cooperation is therefore a key issue for good control performance, and study on team cooperation processes is an important issue.

We have been studying team cooperation processes in en-route ATC based on ethnographic field observation, and already proposed a cognitive model of team cooperation in en-route ATC as shown in Fig. 1 (Furuta et al., 2009; Soraji et al., 2010). In this model, establishment of Team Situation Awareness (TSA) on the air traffic is a key process for smooth cooperation. TSA here can be defined as a combination of individual situation awareness (Endsley, 1995) and mutual beliefs on it (Shu & Furuta, 2005). Once TSA has been established, individual tasks will be planned and executed almost implicitly, and TSA almost determines decision by ATCOs. The cognitive processes of a controller team after TSA acquisition are well described by a distributed version of the recognition-primed decision model (Klein, 1997).

© 2012 Furuta et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Furuta et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

70 Advances in Air Navigation Services

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 71

beliefs on his/her partner's cognition, and it reflects partner's first layer. The third layer describes ones beliefs on partner's beliefs on his/her own cognition, which are the self-image through the partner. Information not only on the environment but also on partner's cognition is used in team cooperation, and all of them are described in the layers of beliefs in

Verification

A's mind B's mind

Complementing Communication Mental simulation

Each belief item in this structure is obtained through internal or interpersonal interactions as well as through perception directly from the environment. There are four types of interactions: verbal and nonverbal communication, mental simulation (inference), complementing (presumption), and verification (consistency check). Communication is a process to transfer some belief item from one person to another by explicit utterance or observation. Mental simulation is a process to derive new belief items from some others within the same MBM layer by inference. In complementing, some belief item will be copied from one MBM layer to another within the same person. Verification is comparison of belief items between different MBM layers within the same person to check consistency among

Manifestation styles of interpersonal interactions are threefold: informing, query and answer (Q&A), and observation. Informing is an interaction to transfer some information from one person to another. This interaction is carried out by writing into a flight data strip as well as by verbal communication in ATC. Q&A is a combination of requesting information and corresponding reply to the request. This interaction is carried out almost verbally and it is used primarily for verification. Query is also used for the purpose of informing, and no replies are returned often in such a case. Observation is guessing partner's cognitive processes or mental states from his/her performance. The observed

 B's beliefs on A's beliefs on B's cognition

 B's beliefs on A's cognition

B's cognition

MBM.

1st layer

2nd layer

mutual beliefs.

3rd layer A's beliefs on

**Figure 2.** Mutual Belief Model and interaction schemes

obtains a belief that he/she is being observed in most cases.

 B's beliefs on A's cognition

 A's beliefs on B's cognition

A's cognition

**Figure 1.** Cognitive model of team cooperative processes

Though the proposed cognitive model can explain the primary features of team cooperation in en-route ATC, it is still static and cannot explain dynamics of the processes. The aim of this research is to study detailed cognitive processes and communication strategies for establishing TSA using computer simulation. Computer simulation has been a useful tool in cognitive systems engineering for validation and sophistication of cognitive models (e.g., Furuta & Kondo, 1993; Cacciabue & Cojazzi, 1995; Cacciabue, 1998; Chang & Mosleh, 2007; NASA, 2011), because no ambiguities are allowed in coding executable computer programs for simulation. Interaction schemes and communication strategies for establishing TSA were discussed in our previous study, and computer simulation is a good approach also to reveal how ATCOs organize these schemes and strategies in their actual field settings.

#### **2. Theoretical backgrounds**

This study as well as our previous study relies on the Mutual Belief Model (MBM) as the theoretical basis. MBM is represented as a three-layered structure of items believed by team members, and it is a framework to explain how a team of individuals can coordinate their tasks smoothly to achieve shared team goals (Kanno, 2007). It is premised in MBM that there is a layered structure of human beliefs for all cognitive constructs and that establishing its consistency enables one to cooperate his/her partners. Figure 2 illustrates MBM for a team with two members, A and B. The layered structure theoretically repeats ad infinitum, but considering the first three layers will be enough in the real situation.

The first layer is the place to describe ones own cognition: what are perceived, recognized, believed, predicted, intended, planed, and so on. The second layer is for describing ones beliefs on his/her partner's cognition, and it reflects partner's first layer. The third layer describes ones beliefs on partner's beliefs on his/her own cognition, which are the self-image through the partner. Information not only on the environment but also on partner's cognition is used in team cooperation, and all of them are described in the layers of beliefs in MBM.

**Figure 2.** Mutual Belief Model and interaction schemes

70 Advances in Air Navigation Services

Interpersonal processes

Personal processes

Acquisition of individual situation awareness

**Figure 1.** Cognitive model of team cooperative processes

**2. Theoretical backgrounds** 

Situation of target sector

> Belief acquisition Informing Q & A Observation

Mental simulation using knowledge and experience

to Partner from Partner

Though the proposed cognitive model can explain the primary features of team cooperation in en-route ATC, it is still static and cannot explain dynamics of the processes. The aim of this research is to study detailed cognitive processes and communication strategies for establishing TSA using computer simulation. Computer simulation has been a useful tool in cognitive systems engineering for validation and sophistication of cognitive models (e.g., Furuta & Kondo, 1993; Cacciabue & Cojazzi, 1995; Cacciabue, 1998; Chang & Mosleh, 2007; NASA, 2011), because no ambiguities are allowed in coding executable computer programs for simulation. Interaction schemes and communication strategies for establishing TSA were discussed in our previous study, and computer simulation is a good approach also to reveal

This study as well as our previous study relies on the Mutual Belief Model (MBM) as the theoretical basis. MBM is represented as a three-layered structure of items believed by team members, and it is a framework to explain how a team of individuals can coordinate their tasks smoothly to achieve shared team goals (Kanno, 2007). It is premised in MBM that there is a layered structure of human beliefs for all cognitive constructs and that establishing its consistency enables one to cooperate his/her partners. Figure 2 illustrates MBM for a team with two members, A and B. The layered structure theoretically repeats ad infinitum, but

The first layer is the place to describe ones own cognition: what are perceived, recognized, believed, predicted, intended, planed, and so on. The second layer is for describing ones

how ATCOs organize these schemes and strategies in their actual field settings.

considering the first three layers will be enough in the real situation.

Choosing scheme of interaction

Acquisition of TSA

Decision making based on TSA

Execution of control instruction or coordination

> Task model (Role model)

Detection and repair of insufficiency / inconsistency in TSA Complementing Verification

> Each belief item in this structure is obtained through internal or interpersonal interactions as well as through perception directly from the environment. There are four types of interactions: verbal and nonverbal communication, mental simulation (inference), complementing (presumption), and verification (consistency check). Communication is a process to transfer some belief item from one person to another by explicit utterance or observation. Mental simulation is a process to derive new belief items from some others within the same MBM layer by inference. In complementing, some belief item will be copied from one MBM layer to another within the same person. Verification is comparison of belief items between different MBM layers within the same person to check consistency among mutual beliefs.

> Manifestation styles of interpersonal interactions are threefold: informing, query and answer (Q&A), and observation. Informing is an interaction to transfer some information from one person to another. This interaction is carried out by writing into a flight data strip as well as by verbal communication in ATC. Q&A is a combination of requesting information and corresponding reply to the request. This interaction is carried out almost verbally and it is used primarily for verification. Query is also used for the purpose of informing, and no replies are returned often in such a case. Observation is guessing partner's cognitive processes or mental states from his/her performance. The observed obtains a belief that he/she is being observed in most cases.

### **3. Review of field observation**

#### **3.1. Field observation**

This work is based on the same data from the field observation of our previous work [1, 2]. The field observation was performed at the Tokyo Area Control Centre from 7 to 11 of May, 2007 during time periods of around three hours a day with relatively heavy traffic in the daytime so that the traffic imposed a certain level of workload on ATCOs. Different RC-CC pairs who were on a shift for the target sector were observed, but neither other sectors nor shift supervisors were observed. The target of observation was a sector called "Kanto-north" (T03), which spreads over the northern area of Tokyo. A lot of air traffic that departs from and arrives at two hub airports, the Tokyo International Airport (HND) and the Narita International Airport (NRT), smaller airports, and Air Force Bases (AFBs) passes through this sector.

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 73

transfer information from the speaker to the listener. If the speaker talks his/her own cognition, the information is transferred from speaker's first layer to listener's second layer. If the speaker talks his/her belief on listener's cognition, the information is transferred from speaker's second layer to listener's third layer. Observation is a nonverbal communication of observing partner's behavior to get belief on his/her belief in observer's second layer. Completion is a subcategory of complementing in the previous chapter. It is a process to accept what is believed by ones partner as his/her own belief; it is achieved by copying the belief item from the second to the first layer. Assumption is another subcategory of complementing, where one assumes that what he/she said to the partner is accepted. It is achieved by copying the belief item from the third to the second layer. Completion and assumption are mirroring processes between the speaker and the listener. Inference is the same process as the mental simulation in the previous chapter that derives new belief items from some others within the same layer. Query is a request of transmission for

Interaction Type Origin Destination

In order to exactly describe interactions between ATCOs including internal cognitive processes, we defined several classes of mental constructs, which are basic and unit description of beliefs in ATCO's MBM structure. Using the mental constructs, the results of cognitive task analysis of our previous work were transcribed down in a formal expression similar to predicate. The following are primary mental constructs and their meanings.

Environment 1st layer 2nd layer 1st layer 2nd layer 3rd layer 1st layer 2nd layer 3rd layer

1st layer 2nd layer 3rd layer 2nd layer 1st layer 2nd layer 1st layer 2nd layer 3rd layer

verification of the consistency of belief items in different layers.

Personal Interpersonal

Interpersonal Personal Personal Personal

Perception Transmission

Observation Completion Assumption Inference

**Table 1.** Classification of interactions

*focus(Aircraft1, Aircraft2, …)* 

*priority(Aircraft1, Aircraft2, …)* 

*execute(Action, Arg1, Arg2, …)* 

*instruction(Aircraft, Parameter, Value)* 

*constraint(Aircraft, Parameter, Value)* 

**3.3. Visualizing analysis results of field data** 

 *Attend and handle Aircraft1, Arcraft2, and so on in a group.* 

 *Execute an Action with arguments of Arg1, Arg2, and so on.* 

 *The priority among Aircraft1, Aircraft2, and so on is in this order.* 

 *Give Aircraft an instruction to keep Parameter at/below/above Value.* 

 *Consider a constraint for Aircraft to keep Parameter at/below/above Value.* 

Behavior of ATCOs was vide-recorded with two home video cameras, and another one recorded the radar screen of an auxiliary console where the same screen image was displayed that RCs were monitoring. An IC recorder attached on the controller console above the radar screen was used to record conversation between ATCOs. Flight data and radio communication records were also provided from the control centre. Combined videoaudio records were made from the audio data and the video data of the radar screen synchronizing their time stamps. Radio communication and conversation between ATCOs were then transcribed, and speakers and listeners of conversation were identified. Actions of ATCOs were next recognized from the video data and added to the transcribed protocol data.

With the help of a rated ATCO, we segmented the protocol data by the basic unit of ATC instruction, clarified relations between segments, and identified expert knowledge and judgment behind them. Distributed cognitive processes among ATCOs were then inferred and reconstructed from the field data following the classification framework of MBM interactions. First we did not classify utterances themselves into particular categories but focused on reconstruction of MBM interactions. When we constructed a team cooperation model afterwards, we considered correspondence between the both and interaction schemes to classify utterances considering manifestation of communication.

#### **3.2. Reclassification of interactions**

The classification of interactions given in the previous chapter is a little ambiguous to describe precise processes of team cooperation; we will modify the classification slightly. Since interactions are basically information transfer from somewhere to somewhere, they are classifiable by the origin and the destination of information transfer. Table 1 summarizes the new classification.

Perception is a special interaction that the perceiver acquires information actively or passively from his/her working environment rather than another person. The acquired information is added into the first layer of the perceiver. Transmission (informing) is a verbal communication to transfer information from the speaker to the listener. If the speaker talks his/her own cognition, the information is transferred from speaker's first layer to listener's second layer. If the speaker talks his/her belief on listener's cognition, the information is transferred from speaker's second layer to listener's third layer. Observation is a nonverbal communication of observing partner's behavior to get belief on his/her belief in observer's second layer. Completion is a subcategory of complementing in the previous chapter. It is a process to accept what is believed by ones partner as his/her own belief; it is achieved by copying the belief item from the second to the first layer. Assumption is another subcategory of complementing, where one assumes that what he/she said to the partner is accepted. It is achieved by copying the belief item from the third to the second layer. Completion and assumption are mirroring processes between the speaker and the listener. Inference is the same process as the mental simulation in the previous chapter that derives new belief items from some others within the same layer. Query is a request of transmission for verification of the consistency of belief items in different layers.


**Table 1.** Classification of interactions

72 Advances in Air Navigation Services

**3.1. Field observation** 

this sector.

protocol data.

**3. Review of field observation** 

This work is based on the same data from the field observation of our previous work [1, 2]. The field observation was performed at the Tokyo Area Control Centre from 7 to 11 of May, 2007 during time periods of around three hours a day with relatively heavy traffic in the daytime so that the traffic imposed a certain level of workload on ATCOs. Different RC-CC pairs who were on a shift for the target sector were observed, but neither other sectors nor shift supervisors were observed. The target of observation was a sector called "Kanto-north" (T03), which spreads over the northern area of Tokyo. A lot of air traffic that departs from and arrives at two hub airports, the Tokyo International Airport (HND) and the Narita International Airport (NRT), smaller airports, and Air Force Bases (AFBs) passes through

Behavior of ATCOs was vide-recorded with two home video cameras, and another one recorded the radar screen of an auxiliary console where the same screen image was displayed that RCs were monitoring. An IC recorder attached on the controller console above the radar screen was used to record conversation between ATCOs. Flight data and radio communication records were also provided from the control centre. Combined videoaudio records were made from the audio data and the video data of the radar screen synchronizing their time stamps. Radio communication and conversation between ATCOs were then transcribed, and speakers and listeners of conversation were identified. Actions of ATCOs were next recognized from the video data and added to the transcribed

With the help of a rated ATCO, we segmented the protocol data by the basic unit of ATC instruction, clarified relations between segments, and identified expert knowledge and judgment behind them. Distributed cognitive processes among ATCOs were then inferred and reconstructed from the field data following the classification framework of MBM interactions. First we did not classify utterances themselves into particular categories but focused on reconstruction of MBM interactions. When we constructed a team cooperation model afterwards, we considered correspondence between the both and interaction schemes

The classification of interactions given in the previous chapter is a little ambiguous to describe precise processes of team cooperation; we will modify the classification slightly. Since interactions are basically information transfer from somewhere to somewhere, they are classifiable by the origin and the destination of information transfer. Table 1 summarizes

Perception is a special interaction that the perceiver acquires information actively or passively from his/her working environment rather than another person. The acquired information is added into the first layer of the perceiver. Transmission (informing) is a verbal communication to

to classify utterances considering manifestation of communication.

**3.2. Reclassification of interactions** 

the new classification.

#### **3.3. Visualizing analysis results of field data**

In order to exactly describe interactions between ATCOs including internal cognitive processes, we defined several classes of mental constructs, which are basic and unit description of beliefs in ATCO's MBM structure. Using the mental constructs, the results of cognitive task analysis of our previous work were transcribed down in a formal expression similar to predicate. The following are primary mental constructs and their meanings.

*focus(Aircraft1, Aircraft2, …) Attend and handle Aircraft1, Arcraft2, and so on in a group. priority(Aircraft1, Aircraft2, …) The priority among Aircraft1, Aircraft2, and so on is in this order. instruction(Aircraft, Parameter, Value) Give Aircraft an instruction to keep Parameter at/below/above Value. constraint(Aircraft, Parameter, Value) Consider a constraint for Aircraft to keep Parameter at/below/above Value. execute(Action, Arg1, Arg2, …) Execute an Action with arguments of Arg1, Arg2, and so on.* 

#### 74 Advances in Air Navigation Services

Describing analysis results in such a representation enables one to compare them easily with those of computer simulation.

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 75

item is transferred to the second layer of CC by transmission or observation in the next moment. CC then copies the item to the first layer by completion and informs RC of his/her belief to make another copy of the item in RC's second layer. Having finished all these steps successfully, the team has copies on the same belief item in both the first and the second layer, and they form sound TSA. The complementary pattern exists that starts from

Finally, after having acquired TSA, ATCOs will start interactions for deepening their

Computer simulation of interactions between RC and CC has been developed. Each member of the team is modelled as an agent in the simulation system: RC agent and CC agent. The both agents repeat choosing and executing a cognitive task to modify beliefs by referring to

Each agent first looks at the mutual belief module and create a list of cognitive tasks applicable to the current situation. A cognitive task is represented as a production rule, which consists of the condition part and the action part. The general rules related to basic interactions for establishing TSA are hard-coded in the agents, and their implementation is

thoughts based on the acquired TSA like considering strategies for the focused issue.

its own mutual belief module. Figure 4 illustrates the flow of simulation.

perception by CC.

explained below.

**Figure 4.** Flow of simulation

**4. Simulation model** 

**4.1. Flow of simulation** 

#### **3.4. Features of team interactions between ATCOs**

In this study, we first developed a tool for visualizing the analysis results of the previous work on team interactions for detailed review. The same tool is also used later to visualize simulation results. The tool reads a file of formal descriptions of analysis or simulation results, and chronologically visualizes interactions between ATCOs by animating moves of information between MBM layers. Consequently, four features of team interactions have been found from the visualization. These features provide valuable hints and justifications in constructing a simulation model of team cooperation processes.

The first feature is the dominance of RC, which means that RC's cognition often starts a sequence of team cognitive processes. Perception by RC and succeeding transmission of the perceived information are frequently observed in the field data. Observation by CC on RC's behaviour is substitutive for transmission by RC in the above. Information transfer from RC to CC is more frequent than that in the opposite direction. Questioning by RC to CC is sometimes observed, but it is primarily for informing CC of RC's thought to request CC's support rather than literally asking CC's thought. Team cognitive processes are thereby paced by RC's cognition.

**Figure 3.** Interaction pattern for efficient TSA development

Secondly a recent topic is likely to be focused on in team interactions and deliberated further. It is the so-called recency effect, which is frequently observed in general cognitive processes.

The third finding is that an efficient pattern of TSA development often appears. Figure 3 shows a case where the pattern starts from perception of something by RC. The perceived item is transferred to the second layer of CC by transmission or observation in the next moment. CC then copies the item to the first layer by completion and informs RC of his/her belief to make another copy of the item in RC's second layer. Having finished all these steps successfully, the team has copies on the same belief item in both the first and the second layer, and they form sound TSA. The complementary pattern exists that starts from perception by CC.

Finally, after having acquired TSA, ATCOs will start interactions for deepening their thoughts based on the acquired TSA like considering strategies for the focused issue.

### **4. Simulation model**

74 Advances in Air Navigation Services

paced by RC's cognition.

processes.

those of computer simulation.

**3.4. Features of team interactions between ATCOs** 

in constructing a simulation model of team cooperation processes.

**Figure 3.** Interaction pattern for efficient TSA development

Describing analysis results in such a representation enables one to compare them easily with

In this study, we first developed a tool for visualizing the analysis results of the previous work on team interactions for detailed review. The same tool is also used later to visualize simulation results. The tool reads a file of formal descriptions of analysis or simulation results, and chronologically visualizes interactions between ATCOs by animating moves of information between MBM layers. Consequently, four features of team interactions have been found from the visualization. These features provide valuable hints and justifications

The first feature is the dominance of RC, which means that RC's cognition often starts a sequence of team cognitive processes. Perception by RC and succeeding transmission of the perceived information are frequently observed in the field data. Observation by CC on RC's behaviour is substitutive for transmission by RC in the above. Information transfer from RC to CC is more frequent than that in the opposite direction. Questioning by RC to CC is sometimes observed, but it is primarily for informing CC of RC's thought to request CC's support rather than literally asking CC's thought. Team cognitive processes are thereby

Secondly a recent topic is likely to be focused on in team interactions and deliberated further. It is the so-called recency effect, which is frequently observed in general cognitive

The third finding is that an efficient pattern of TSA development often appears. Figure 3 shows a case where the pattern starts from perception of something by RC. The perceived

#### **4.1. Flow of simulation**

Computer simulation of interactions between RC and CC has been developed. Each member of the team is modelled as an agent in the simulation system: RC agent and CC agent. The both agents repeat choosing and executing a cognitive task to modify beliefs by referring to its own mutual belief module. Figure 4 illustrates the flow of simulation.

Each agent first looks at the mutual belief module and create a list of cognitive tasks applicable to the current situation. A cognitive task is represented as a production rule, which consists of the condition part and the action part. The general rules related to basic interactions for establishing TSA are hard-coded in the agents, and their implementation is explained below.

**Figure 4.** Flow of simulation

#### 76 Advances in Air Navigation Services

**Observation (obs):** If an agent executes any task visible to the partner agent, the latter will see the task and create the corresponding belief item in its second layer.

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 77

Bonus for efficient TSA development Bonus for efficient TSA development Bonus for efficient TSA development Bonus for efficient TSA development

Bonus for the recency effect

Basic score Basic score Basic score Basic score Basic score Basic score Basic score

data, it is labelled "perfect match." An internal task that is unobservable from the outside of an individual but predictable will be labelled "predictive match," if it is included in the simulation results. "Essential match" means that a pair of tasks in the observation data and the simulation results has the same effects on mutual beliefs, though they are different in

A measure of TSA appropriateness is completeness, which is defined as the ratio of belief items shared by the two corresponding belief layers of a team over the total number of belief items in the mirrored layer of the pair [4]. In these case studies, completeness of the second MBM layers was evaluated, and one should notice that just belief items relevant to related aircrafts were considered in simulation. Consequently, the results are more extreme than in the real situation, where ATCOs conceive belief items irrelevant to related aircrafts. Another measure, soundness, was ignored in this study, because no error mechanism was considered in the simulation model of this work and the soundness of TSA will be almost 100% in any

In the first case (Fig. 5), the controller team dealt with a situation where three aircrafts bound for NRT from the northwest, the first one of which is denoted ac1 below, and another one from the north, ac2, was entering the sector almost at the same time. In addition, a warplane, ac3, were approaching from the southwest to cross over the flight path towards NRT, and it might interfere one of the four aircrafts. The controller team discussed and decided the order of aircrafts descending toward NRT while managing separation of the

Table 3 compares simulated and observed interactions in this case. Items printed in italics were given in the simulation scenario and those in bolds show matched items between

Task type or task sequence Score Explanation

appearance.

case.

Perception (per) Transmission (trn) Observation (obs) Completion (com) Assumption (ass) Inference (inf) Execution (exe)

Transmission Assumption Perception Transmission Inference Transmission Transmission Completion

**5.1. Case 1 (P.M, may 7, 2007)** 

warplane from the others.

simulation and observation.

**Table 2.** Scores for prioritizing cognitive tasks

**Completion (com):** If any belief item contained in the second layer of an agent is missing in the first layer, the agent tries to copy the item into the first layer.

**Assumption (ass):** If any belief item contained in the third layer of an agent is missing in the second layer, the agent tries to copy the item into the second layer.

**Inference (inf):** A production rule fetched from the rule base is applied, if the rule is applicable to the present situation. Since this simulation exclusively focuses on team cooperation processes, it does not deal with the domain expertise specific to ATC explicitly. Expert judgment and inference are modelled as simple application of predefined rules that are very specific to each simulation case.

**Perception (per):** All perception tasks are predefined in the simulation scenario and they are triggered by time as interruption. Simulation scenarios are to be set up based on the field observation data, if simulation is to be done for the situations observed in the field.

**Transmission (trn):** If any belief item in agent's first layer does not exist in its second layer, the agent attempts to inform the item to the partner agent, i.e., tries to copy the item into the second layer of the partner agent. In execution of transmission, the transmitter agent also copies the item also into the third layer of its own, assuming that transmission is successfully done.

There is another kind of cognitive task, execution (exe), which is to execute some control action like hand-off, point-out, or giving a control instruction to a pilot. Execution is triggered as a result of rule application by inference. Query (qry) is not yet implemented in the present simulation model.

#### **4.2. Prioritization of cognitive tasks**

The agents next prioritize cognitive tasks in the created task lists. Cognitive tasks are scored referring not only to the basic score predefined for each task type but also to the past records of simulation hold in the memory models. When a cognitive task is triggered, its basic score shown in Table 2 is first given. The recency effect observed in the field data can be taken into consideration by adding a bonus if the task is related to the belief item created in the previous simulation step. If a particular sequence appeared between the successive tasks, another bonus will be added, which promotes the interaction pattern for efficient TSA development discussed in the previous chapter. After scoring, the task with the highest score is chosen and executed.

The scores have been adjusted so that simulation can well replicate the interactions observed in the actual ATCOs, and Table 2 lists the final scores.

The degree of match between the analysis of observation data and the simulation results was evaluated in three levels. If simulation could predict exactly a task appeared in the field data, it is labelled "perfect match." An internal task that is unobservable from the outside of an individual but predictable will be labelled "predictive match," if it is included in the simulation results. "Essential match" means that a pair of tasks in the observation data and the simulation results has the same effects on mutual beliefs, though they are different in appearance.


**Table 2.** Scores for prioritizing cognitive tasks

76 Advances in Air Navigation Services

are very specific to each simulation case.

successfully done.

the present simulation model.

score is chosen and executed.

in the actual ATCOs, and Table 2 lists the final scores.

**4.2. Prioritization of cognitive tasks** 

**Observation (obs):** If an agent executes any task visible to the partner agent, the latter will

**Completion (com):** If any belief item contained in the second layer of an agent is missing in

**Assumption (ass):** If any belief item contained in the third layer of an agent is missing in the

**Inference (inf):** A production rule fetched from the rule base is applied, if the rule is applicable to the present situation. Since this simulation exclusively focuses on team cooperation processes, it does not deal with the domain expertise specific to ATC explicitly. Expert judgment and inference are modelled as simple application of predefined rules that

**Perception (per):** All perception tasks are predefined in the simulation scenario and they are triggered by time as interruption. Simulation scenarios are to be set up based on the field

**Transmission (trn):** If any belief item in agent's first layer does not exist in its second layer, the agent attempts to inform the item to the partner agent, i.e., tries to copy the item into the second layer of the partner agent. In execution of transmission, the transmitter agent also copies the item also into the third layer of its own, assuming that transmission is

There is another kind of cognitive task, execution (exe), which is to execute some control action like hand-off, point-out, or giving a control instruction to a pilot. Execution is triggered as a result of rule application by inference. Query (qry) is not yet implemented in

The agents next prioritize cognitive tasks in the created task lists. Cognitive tasks are scored referring not only to the basic score predefined for each task type but also to the past records of simulation hold in the memory models. When a cognitive task is triggered, its basic score shown in Table 2 is first given. The recency effect observed in the field data can be taken into consideration by adding a bonus if the task is related to the belief item created in the previous simulation step. If a particular sequence appeared between the successive tasks, another bonus will be added, which promotes the interaction pattern for efficient TSA development discussed in the previous chapter. After scoring, the task with the highest

The scores have been adjusted so that simulation can well replicate the interactions observed

The degree of match between the analysis of observation data and the simulation results was evaluated in three levels. If simulation could predict exactly a task appeared in the field

observation data, if simulation is to be done for the situations observed in the field.

see the task and create the corresponding belief item in its second layer.

the first layer, the agent tries to copy the item into the first layer.

second layer, the agent tries to copy the item into the second layer.

A measure of TSA appropriateness is completeness, which is defined as the ratio of belief items shared by the two corresponding belief layers of a team over the total number of belief items in the mirrored layer of the pair [4]. In these case studies, completeness of the second MBM layers was evaluated, and one should notice that just belief items relevant to related aircrafts were considered in simulation. Consequently, the results are more extreme than in the real situation, where ATCOs conceive belief items irrelevant to related aircrafts. Another measure, soundness, was ignored in this study, because no error mechanism was considered in the simulation model of this work and the soundness of TSA will be almost 100% in any case.

#### **5.1. Case 1 (P.M, may 7, 2007)**

In the first case (Fig. 5), the controller team dealt with a situation where three aircrafts bound for NRT from the northwest, the first one of which is denoted ac1 below, and another one from the north, ac2, was entering the sector almost at the same time. In addition, a warplane, ac3, were approaching from the southwest to cross over the flight path towards NRT, and it might interfere one of the four aircrafts. The controller team discussed and decided the order of aircrafts descending toward NRT while managing separation of the warplane from the others.

Table 3 compares simulated and observed interactions in this case. Items printed in italics were given in the simulation scenario and those in bolds show matched items between simulation and observation.

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 79

*RC: per - focus(ac1, ac2)*  **CC: obs - focus(ac1, ac2)\* RC: inf - priority(ac1, ac2)+ CC: qry - priority(ac1, ac2)^ CC: com - focus(ac1, ac2)+ CC: com - priority(ac1, ac2)+**  CC: trn - priority(ac1, ac2)

*RC: per - focus(ac1, ac2, ac3)*  **CC: obs - focus(ac1, ac2, ac3)\*** 

**CC: com - focus(ac1, ac2, ac3)+** 

**CC: inf - instruct(ac3, alt, down)+ CC: trn - instruct(ac3, alt, down)\* RC: com - instruct(ac3, alt, down)+** 

**RC: inf - instruct(ac3, alt, 130)+ RC: trn - instruct(ac3, alt, 130)\* CC: com - instruct(ac3, alt, 130)+** CC: trn - instruct(ac3, alt, 130) *CC: per - constraint(ac3, alt, 170)*  **CC: trn - constraint(ac3, alt, 170)\*** 

*RC: per - focus(ac1', ac2')* 

**RC: qry - focus(ac1', ac2')^** 

**CC: com - focus(ac1', ac2')+** 

**CC: inf - priority(ac2, ac1)+ CC: trn - priority(ac2, ac1)\* RC: com - priority(ac2, ac1)+ RC: com - constraint(ac3, alt, 170)+**

Simulation Observation

14:07:40

14:07:47

14:08:09

14:08:22

14:08:48

14:08:57 14:09:51

14:10:01

14:10:27 14:10:55

Step Actor and action Time Actor and action

0

*RC: per - focus(ac1, ac2)*  **CC: obs - focus(ac1, ac2)\***  RC: trn - focus(ac1,ac2) **CC: com - focus(ac1, ac2)+**  RC: ass - focus(ac1, ac2) **RC: inf - priority(ac1, ac2)+ RC: trn - priority(ac1, ac2)^**  RC: ass - priority(ac1, ac2) **CC: com - priority(ac1, ac2)+**  *RC: per - focus(ac1, ac2, ac3)*  **CC: obs - focus(ac1, ac2, ac3)\***  RC: trn - focus(ac1, ac2, ac3) **CC: com - focus(ac1, ac2, ac3)+**  RC: ass - focus(ac1, ac2, ac3) **CC: inf - instruct(ac3, alt down)+ CC: trn - instruct(ac3, alt, down)\* RC: com - instruct(ac3, alt, down)+**  CC: ass - instruct(ac3, alt, down) **RC: inf - instruct(ac3, alt, 130)+ RC: trn - instruct(ac3, alt, 130)\***  AC: ass - instruct(ac3, alt, 130) **CC: com - instruct(ac3, alt, 130)+**  *CC: per - constraint(ac3, alt, 170)*  **CC: trn - constraint(ac3, alt, 170)\*** 

*RC: per - focus(ac1', ac2')* 

*Italics*: Items given in the simulation scenario.

**Bolds**: Items matched between simulation and observation. (\* Perfect match + Predictive match ^ Essential match)

**Table 3.** Comparison of simulated and observed interactions in Case 1

CC: ass - constraint(ac3, alt, 170) **RC: trn - focus(ac1', ac2')^**  AC: ass - focus(ac1', ac2') **CC: com - focus(ac1', ac2')+ RC: com -constraint(ac3, alt, 170)+ CC: inf - priority(ac2, ac1)+ CC: trn - priority(ac2, ac1)\* RC: com - priority(ac2, ac1)+**  CC: ass - priority(ac2, ac1)

1

6

7

8

9 10

11 12 13

17 18 19

20 21

22

23 24

**Figure 5.** Situation of traffic in Case 1

Having watched the radar screen, RC judged the priority between ac1 and ac2 as ac1 should go first. CC recognized that RC was considering the order of ac1 and ac2 through observation of RC, CC asked a question on RC's judgment in the field, while RC transmitted its judgment explicitly to CC in simulation. RC then recognized that ac3 would interfere the other aircrafts, and CC proposed to instruct ac3 to go down in advance. RC agreed and proposed a particular altitude of descent as FL130, but CC came to know that ac3 should be kept above FL170 as a result of coordination with Hyakuri AFB control. Meanwhile, RC continued monitoring the positions of aircrafts, and CC observed these actions. CC proposed to change the order of aircrafts as ac2 should go first, and transmitted this judgment to RC.

Among 26 interactions that were extracted from the field data, four items were used to setup the simulation scenario, and simulation successfully replicated 18 items: 81.8% (18/22) of interactions observed in the field data. The order of appearance of interactions was almost the same. Two queries observed in the field were simulated as transmissions, and two transmissions by CC were not simulated. Four transmissions by RC predicted by simulation were not executed actually, and no assumptions were observable in the field.

The assessment result of TSA is shown in Fig. 6. Completeness of both RC and CC sometimes degrades, when one of them perceived new information or generated new belief items by inference. It was however recovered soon by verbal communication. TSA completeness of CC was slightly better than that of RC, but difference was very small, because they kept close verbal communication for deciding the priority of related aircrafts.


*Italics*: Items given in the simulation scenario.

78 Advances in Air Navigation Services

 ac3

for Hyakuri

**Figure 5.** Situation of traffic in Case 1

judgment to RC.

field.

aircrafts.

 

ac1

Having watched the radar screen, RC judged the priority between ac1 and ac2 as ac1 should go first. CC recognized that RC was considering the order of ac1 and ac2 through observation of RC, CC asked a question on RC's judgment in the field, while RC transmitted its judgment explicitly to CC in simulation. RC then recognized that ac3 would interfere the other aircrafts, and CC proposed to instruct ac3 to go down in advance. RC agreed and proposed a particular altitude of descent as FL130, but CC came to know that ac3 should be kept above FL170 as a result of coordination with Hyakuri AFB control. Meanwhile, RC continued monitoring the positions of aircrafts, and CC observed these actions. CC proposed to change the order of aircrafts as ac2 should go first, and transmitted this

Among 26 interactions that were extracted from the field data, four items were used to setup the simulation scenario, and simulation successfully replicated 18 items: 81.8% (18/22) of interactions observed in the field data. The order of appearance of interactions was almost the same. Two queries observed in the field were simulated as transmissions, and two transmissions by CC were not simulated. Four transmissions by RC predicted by simulation were not executed actually, and no assumptions were observable in the

The assessment result of TSA is shown in Fig. 6. Completeness of both RC and CC sometimes degrades, when one of them perceived new information or generated new belief items by inference. It was however recovered soon by verbal communication. TSA completeness of CC was slightly better than that of RC, but difference was very small, because they kept close verbal communication for deciding the priority of related

for NRT for NRT

ac2

T03

**Bolds**: Items matched between simulation and observation.

(\* Perfect match + Predictive match ^ Essential match)

**Table 3.** Comparison of simulated and observed interactions in Case 1

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 81

*RC: per - point\_out(ac1, HND)*  **RC: trn - point\_out(ac1, HND)\*** 

*CC: per - focus(ac1, ac2)*  **CC: inf - coordinate(ac2, alt)+ CC: trn - coordinate(ac2, alt)\* CC: trn - focus(ac1, ac2)\* RC: com - focus(ac1, ac2)+**  RC: trn - focus(ac1, ac2) **RC: com - coordinate(ac2, alt)+ CC: exe - coordinate(ac2, alt)\***  CC: trn - coordinate(ac2, alt, 130) RC: com - coordinate(ac2, alt, 130)

*RC: per - enter(ac1, NRT)*  **RC: trn - enter(ac1, NRT)\* CC: com - enter(ac1, NRT)+ CC: inf - point\_out(ac1, NRT)+ CC: trn - point\_out(ac1, NRT)\* RC: com - point\_out(ac1, NRT)+** 

**RC: inf - point\_out(ac1, NRT, C)+ RC: qry - point\_out(ac1, NRT, C)^** 

**CC: inf - point\_out(ac1, NRT, D)+ CC: trn - point\_out(ac1, NRT, D)\* RC: com - point\_out(ac1, NRT, D)+**  RC: trn - point\_out(ac1, NRT, D)

**CC: com - point\_out(ac1, HND)+**

ac2 intended to lower ac2 to fly below ac1. Since ac2 still was under the control of HND, CC made coordination with HND departure control to limit its altitude below FL130. Ac1 would enter the control area of NRT due to its low altitude and CC recognized the necessity of pointing out ac1 to NRT control. CC communicated this possibility to RC. Having heard this communication, RC verified CC by a query with which control of NRT to make coordination: center (C) or departure (D) control. This query in the field data was

Simulation Observation Step Actor and action Time Actor and action

12:08:40

12:09:10 12:09:26

12:09:31

12:12:24

12:12:31

substituted with transmission and completion in simulation.

*RC: per - point\_out(ac1, HND)*  **RC: trn - point\_out(ac1, HND)\***  RC: ass - point\_out(ac1, HND) **CC: com - point\_out(ac1, HND)+** 

*CC: per - focus(ac1, ac2)*  **CC: inf - coordinate(ac2, alt)+ CC: trn - coordinate(ac2, alt)\* RC: com - coordinate(ac2, alt)+**  CC: ass - coordinate(ac2, alt) **CC: exe - coordinate(ac2, alt)\* CC: trn - focus(ac1, ac2)\* RC: com - focus(ac1, ac2)+**  CC: ass - focus(ac1, ac2) *RC: per - enter(ac1, NRT)*  **RC: trn - enter(ac1, NRT)\***  RC: ass - enter(ac1, NRT) **CC: com - enter(ac1, NRT)+ CC: inf - point\_out(ac1, NRT)+ CC: trn - point\_out(ac1, NRT)\* RC: com - point\_out(ac1, NRT)+**  CC: ass - point\_out(ac1, NRT) **RC: inf - point\_out(ac1, NRT, C)+ RC: trn - point\_out(ac1, NRT, C)^**  RC: ass - point\_out(ac1, NRT, C) CC: com - point\_out(ac1, NRT, C) **CC: inf - point\_out(ac1, NRT, D)+ CC: trn - point\_out(ac1, NRT, D)\* RC: com - point\_out(ac1, NRT, D)+**  CC: ass - point\_out(ac1, NRT, D)

0 1 2

7 8 9

10 11 12

13 14 15

16 17 18

19 20 21

*Italics*: Items given in the simulation scenario.

**Bolds**: Items matched between simulation and observation. (\* Perfect match + Predictive match ^ Essential match)

**Table 4.** Comparison of simulated and observed interactions in Case 2

**Figure 6.** Completeness of the second MBM layers in Case 1

#### **5.2. Case 2 (P.M. may 11, 2007)**

In the second case (Fig. 7), an aircraft, ac1, departed from Yokota AFB and flied along the southernmost boundary of Sector T03 eastward to the Pacific Ocean. Many aircrafts departed from HND and NRT passed through the area during this time period; ATCOs had to concern about traffic interference. Another aircraft, ac2, departed from HND was climbing northward. Since the approach areas of NRT, HND and Yokota as well as neighboring sectors overlap in this area, demands for coordination with other sectors are relatively high.

**Figure 7.** Situation of traffic in Case 2

Table 4 compares simulated and observed interactions in Case 2. At first, RC recognized the necessity of pointing out ac1 to HND control. CC who feared interference between ac1 and ac2 intended to lower ac2 to fly below ac1. Since ac2 still was under the control of HND, CC made coordination with HND departure control to limit its altitude below FL130. Ac1 would enter the control area of NRT due to its low altitude and CC recognized the necessity of pointing out ac1 to NRT control. CC communicated this possibility to RC. Having heard this communication, RC verified CC by a query with which control of NRT to make coordination: center (C) or departure (D) control. This query in the field data was substituted with transmission and completion in simulation.


*Italics*: Items given in the simulation scenario.

80 Advances in Air Navigation Services

0

**5.2. Case 2 (P.M. may 11, 2007)** 

relatively high.

**Figure 6.** Completeness of the second MBM layers in Case 1

50

Completeness (%)

100

0 5 10 15 20 25

RC CC

Step

In the second case (Fig. 7), an aircraft, ac1, departed from Yokota AFB and flied along the southernmost boundary of Sector T03 eastward to the Pacific Ocean. Many aircrafts departed from HND and NRT passed through the area during this time period; ATCOs had to concern about traffic interference. Another aircraft, ac2, departed from HND was climbing northward. Since the approach areas of NRT, HND and Yokota as well as neighboring sectors overlap in this area, demands for coordination with other sectors are

Table 4 compares simulated and observed interactions in Case 2. At first, RC recognized the necessity of pointing out ac1 to HND control. CC who feared interference between ac1 and

T03

from HND

ac2 ac1

from Yokota

**Figure 7.** Situation of traffic in Case 2

**Bolds**: Items matched between simulation and observation.

(\* Perfect match + Predictive match ^ Essential match)

**Table 4.** Comparison of simulated and observed interactions in Case 2

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 83

RC's instruction to the pilot, and CC's completeness recovered at the end of simulation. RC, however, did never recognize that CC had conceived a different strategy, and RC's completeness remained low till the end of simulation. Completeness of TSA thereby can be a good measure that exactly shows whether or not the members of a cooperating team share

ac2

GOC

Simulation Observation

14:05:27 14:06:26

14:08:27 14:09:17 14:09:53

14:10:22

Step Actor and action Time Actor and action

from Hyakuri

ac1

T03

*CC: per - focus(ac1, ac2)* 

RC: trn - focus(ac1, ac2) *CC: per - hand\_off(hyakuri, ac1)*  **CC: trn - hand\_off(hyakuri, ac1)\* RC: com - hand\_off(hyakuri, ac1)+ RC: exe - instruction(ac1, dir, S)\* CC: obs - instruction(ac1, dir, S)\* CC: com - instruction(ac1, dir, S)+**  CC: inf - instruction(ac1, dir, S)

**CC: inf - instruction(ac1, dir, N)+ CC: trn - focus(ac1, ac2)\* RC: com - focus(ac1, ac2)+ RC: inf - instruction(ac1, dir, S)+** 

for NRT for HND

situation awareness.

**Figure 9.** Situation of traffic in Case 3

*CC: per - focus(ac1, ac2)* 

*Italics*: Items given in the simulation scenario.

**CC: inf - instruction(ac1, dir, N)+ CC: trn - focus(ac1, ac2)\* RC: com - focus(ac1, ac2)+**  CC: ass - focus(ac1, ac2)

**RC: inf - instruction(ac1, dir S)+**  *CC: per - hand\_off(hyakuri, ac1)*  **CC: trn - hand\_off(hyakuri, ac1)\* RC: com - hand\_off(hyakuri, ac1)+**  CC: ass - hand\_off(hyakuri, ac1) **RC: exe - instruction(ac1, dir, S)\* CC: obs - instruction(ac1, dir, S)\* CC: com - instruction(ac1, dir, S)+** 

**Bolds**: Items matched between simulation and observation. (\* Perfect match + Predictive match ^ Essential match)

**Table 5.** Comparison of simulated and observed interactions in Case 3

9

10

**Figure 8.** Completeness of the second MBM layers in Case 2

Among 25 interactions extracted from the field data, three items were predefined in the simulation scenario, and simulation could predict 17 items, 77.3% (17/22) of the observed. The query observed in the field was substituted with transmission and completion, and no assumptions were observable in the field. Three transmissions and one completion were not simulated.

Figure 8 shows completeness of TSA in Case 2. Dominance of RC appears here, that means, RC leads team cognitive processes and CC follows by eagerly obtaining mutual beliefs on them. Consequently, completeness of TSA is a little better for CC than RC.

#### **5.3. Case 3 (P.M. may 11, 2007)**

The third case shows how RC and CC resolved conflict on control strategy. An aircraft, ac1, which departed from Hyakuri AFB, might interfere with another aircraft, ac2, bound for HND (Fig. 9). CC made coordination with Hyakluri AFB control before hand-off for ac1 to fly north around a fix with identification code GOC. RC, however, instructed ac1 to fly south around GOC. Having heard RC's instruction, CC noticed RC's intention and recovered from the conflicting state of mutual beliefs.

Table 5 compares simulated and observed interactions in the last case. CC expected that RC would give ac1 an instruction to fly north around GOC and made coordination with Hyakuri AFB control for its preparation. RC, however, thought the southern rout is better and gave an instruction to fly south around GOC. CC recognized RC's intention by observation and remedied the conflicting mutual belief.

There were 13 interactions from the field data, two of them were used for the simulation scenario, and 9 of them appeared in the simulation result. The hit rate was 81.2% (9/11). Simulation missed one transmission and one inference, and no assumptions were observable in the field data.

Figure 10 shows the assessment result of TSA completeness. Completeness of TSA was worse for both RC and CC in this case than in the other cases, because there was a conflict on control strategy between RC and CC. Finally CC recognized the conflict by observing RC's instruction to the pilot, and CC's completeness recovered at the end of simulation. RC, however, did never recognize that CC had conceived a different strategy, and RC's completeness remained low till the end of simulation. Completeness of TSA thereby can be a good measure that exactly shows whether or not the members of a cooperating team share situation awareness.

**Figure 9.** Situation of traffic in Case 3

82 Advances in Air Navigation Services

0 0

**5.3. Case 3 (P.M. may 11, 2007)** 

observable in the field data.

**Figure 8.** Completeness of the second MBM layers in Case 2

recovered from the conflicting state of mutual beliefs.

observation and remedied the conflicting mutual belief.

Step

Among 25 interactions extracted from the field data, three items were predefined in the simulation scenario, and simulation could predict 17 items, 77.3% (17/22) of the observed. The query observed in the field was substituted with transmission and completion, and no assumptions were observable in the field. Three transmissions and one completion were not

Figure 8 shows completeness of TSA in Case 2. Dominance of RC appears here, that means, RC leads team cognitive processes and CC follows by eagerly obtaining mutual beliefs on

The third case shows how RC and CC resolved conflict on control strategy. An aircraft, ac1, which departed from Hyakuri AFB, might interfere with another aircraft, ac2, bound for HND (Fig. 9). CC made coordination with Hyakluri AFB control before hand-off for ac1 to fly north around a fix with identification code GOC. RC, however, instructed ac1 to fly south around GOC. Having heard RC's instruction, CC noticed RC's intention and

Table 5 compares simulated and observed interactions in the last case. CC expected that RC would give ac1 an instruction to fly north around GOC and made coordination with Hyakuri AFB control for its preparation. RC, however, thought the southern rout is better and gave an instruction to fly south around GOC. CC recognized RC's intention by

There were 13 interactions from the field data, two of them were used for the simulation scenario, and 9 of them appeared in the simulation result. The hit rate was 81.2% (9/11). Simulation missed one transmission and one inference, and no assumptions were

Figure 10 shows the assessment result of TSA completeness. Completeness of TSA was worse for both RC and CC in this case than in the other cases, because there was a conflict on control strategy between RC and CC. Finally CC recognized the conflict by observing

them. Consequently, completeness of TSA is a little better for CC than RC.

5 10 15 20 25

RC CC

50

Completeness (%)

simulated.

100


*Italics*: Items given in the simulation scenario.

**Bolds**: Items matched between simulation and observation.

(\* Perfect match + Predictive match ^ Essential match)

**Table 5.** Comparison of simulated and observed interactions in Case 3

Simulation of Team Cooperation Processes in En-Route Air Traffic Control 85

The results of analysis on field observation data in en-route ATC were transcribed using predefined cognitive constructs and visualized to reveal typical patterns of team interactions for establishing TSA. Computer simulation then has been developed based on the cognitive model of team cooperation processes of our previous study as well as the revealed typical interaction patterns, and simulation was performed for three cases from the field observation. Simulation could replicate around 80% of ATCOs' interactions and the typical features of interactions observed in the field. The good match indicates that the cognitive model of team cooperation processes proposed in our previous study has a reality. In addition, simulation can explain the cognitive mechanisms of team cooperation processes

Appropriateness of TSA was evaluated using the simulation results evaluating completeness of the second MBM layers. Dominance of RC observed in the field resulted in higher TSA completeness for CC than for RC. It is also shown that TSA completeness degrades when RC or CC obtains new belief items by perception or inference but it is soon recovered by verbal or nonverbal communication. Assessment of TSA by computer

The simulation model of this work still has many limitations though. Firstly, expertise required for en-route ATC was represented as simplistic production rules that are very specific to the cases to be simulated, and this implementation lacks generality. Secondly, this work considers no models of errors, and the interactions generated by this model are normative. Smooth and efficient team cooperation is certainly a key for error-free and safe ATC performance. It is expected but inconclusive from this work, however, if it leads also to a high throughput of the sector, because the observation data were obtained just for time periods with relatively heavy traffic in the daytime and no critical situations happened

Cacciabue, P.C. (1998). *Modeling and Simulation of Human Behaviour in System Control*, ISBN 3-

Cacciabue, P.C. & Cojazzi, G. (1995). An integrated simulation approach for the analysis of pilot-aeroplane interaction, *Control Engineering Practice*, Vol.3, No.2, pp. 257-266, ISSN

simulation is thereby a useful mean to visualize the degree of team cooperation.

during the observation. These issues have been left for future studies.

*Department of Systems Innovation, The University of Tokyo, Japan* 

for not only verbal but also for non-verbal interactions.

**6. Conclusion** 

**Author details** 

Satoru Inoue

**7. References** 

0967-0661

Kazuo Furuta, Kouhei Ohno and Taro Kanno

*Electronic Navigation Research Institute, Japan* 

540-76233-7, Springer, London, UK

**Figure 10.** Completeness of the second MBM layers in Case 3

#### **5.4. Discussion**

Simulation could successfully replicate most of the interactions, around 80% in all cases, observed in the field. Table 6 shows a summary of match between simulated and observed interactions. This good match indicates that the simulation model of team interactions including detailed implementation was appropriate. The details of the model here include the initiation conditions and the prioritization scheme of cognitive tasks, which are required for constructing an executable simulation program. Assumptions, which are completely unobservable from the third party, were simulated but not observed in the field. Human expertise required for en-route ATC is beyond the scope of this simulation, because it is represented as simplistic production rules.

Completeness of CC's second layer outperformed that of RC's second layer in all cases. It means that CC is relatively eager to obtain mutual beliefs compared with RC by following and monitoring RC's actions. In contrast, RC is relatively independent from CC in deciding control actions and their timing. It resulted in dominance of RC observed in the field. Interactions like perception and inference generate new belief items in MBM layers to lower completeness of TSA, but it is usually recovered immediately by some sort of communication. These tasks that lower completeness of TSA, however, contribute to deepen thought on the current issues. It seems a standard style to repeat such a tandem process of deepening thought and establishing TSA in cooperation of ATCOs.


**Table 6.** The degree of match between observed and simulated interactions

#### **6. Conclusion**

84 Advances in Air Navigation Services

0 0

represented as simplistic production rules.

**Figure 10.** Completeness of the second MBM layers in Case 3

Step

Simulation could successfully replicate most of the interactions, around 80% in all cases, observed in the field. Table 6 shows a summary of match between simulated and observed interactions. This good match indicates that the simulation model of team interactions including detailed implementation was appropriate. The details of the model here include the initiation conditions and the prioritization scheme of cognitive tasks, which are required for constructing an executable simulation program. Assumptions, which are completely unobservable from the third party, were simulated but not observed in the field. Human expertise required for en-route ATC is beyond the scope of this simulation, because it is

Completeness of CC's second layer outperformed that of RC's second layer in all cases. It means that CC is relatively eager to obtain mutual beliefs compared with RC by following and monitoring RC's actions. In contrast, RC is relatively independent from CC in deciding control actions and their timing. It resulted in dominance of RC observed in the field. Interactions like perception and inference generate new belief items in MBM layers to lower completeness of TSA, but it is usually recovered immediately by some sort of communication. These tasks that lower completeness of TSA, however, contribute to deepen thought on the current issues. It seems a standard style to repeat such a tandem process of

Number of interactions Case 1 Case 2 Case 3 Simulation scenario 4 3 2 Perfect match 6 7 4 Predictive match 12 10 5 Essential match 2 1 0 Not simulated 2 4 2 Observed interactions 26 25 13

deepening thought and establishing TSA in cooperation of ATCOs.

**Table 6.** The degree of match between observed and simulated interactions

2 4 6 8 10 12

RC CC

50

Completeness (%)

**5.4. Discussion** 

100

The results of analysis on field observation data in en-route ATC were transcribed using predefined cognitive constructs and visualized to reveal typical patterns of team interactions for establishing TSA. Computer simulation then has been developed based on the cognitive model of team cooperation processes of our previous study as well as the revealed typical interaction patterns, and simulation was performed for three cases from the field observation. Simulation could replicate around 80% of ATCOs' interactions and the typical features of interactions observed in the field. The good match indicates that the cognitive model of team cooperation processes proposed in our previous study has a reality. In addition, simulation can explain the cognitive mechanisms of team cooperation processes for not only verbal but also for non-verbal interactions.

Appropriateness of TSA was evaluated using the simulation results evaluating completeness of the second MBM layers. Dominance of RC observed in the field resulted in higher TSA completeness for CC than for RC. It is also shown that TSA completeness degrades when RC or CC obtains new belief items by perception or inference but it is soon recovered by verbal or nonverbal communication. Assessment of TSA by computer simulation is thereby a useful mean to visualize the degree of team cooperation.

The simulation model of this work still has many limitations though. Firstly, expertise required for en-route ATC was represented as simplistic production rules that are very specific to the cases to be simulated, and this implementation lacks generality. Secondly, this work considers no models of errors, and the interactions generated by this model are normative. Smooth and efficient team cooperation is certainly a key for error-free and safe ATC performance. It is expected but inconclusive from this work, however, if it leads also to a high throughput of the sector, because the observation data were obtained just for time periods with relatively heavy traffic in the daytime and no critical situations happened during the observation. These issues have been left for future studies.

#### **Author details**

Kazuo Furuta, Kouhei Ohno and Taro Kanno *Department of Systems Innovation, The University of Tokyo, Japan* 

Satoru Inoue *Electronic Navigation Research Institute, Japan* 

#### **7. References**


Chang, Y.H.J. & Mosleh, A. (2007). Cognitive modelling and dynamic probabilistic simulation of operating crew response to complex system accidents, Part 1: Overview of the IDAC model, *Reliability Engineering and System Safety*, Vol. 92, No.8, pp. 997-1013, ISSN 0951-8320

**Chapter 6** 

© 2012 Arnaldo et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2012 Arnaldo et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**Probability of Potential Collision for Aircraft** 

Collision risk estimation in airspace and mathematical modeling of mid-air collisions have been carried out for over more than 40 years [1]. During this period there has been a development of mathematical models for processes leading to possible collisions of aircraft

B. L. Marks [2] developed the principles in which a collision risk model could be developed in the early 1960s. Marks' work was modified and enhanced by P. Reich [3] and that model, later called the Reich model, has been the basis for many of the important developments in

The Reich model uses information related to the probabilistic distributions of aircraft's lateral and vertical position, traffic flows of the routes, aircraft's relative velocities and aircraft dimensions to generate estimation of collision risk. However, this model does not cover adequately situations where ground controllers monitor the air traffic through radar surveillance and provide tactical instructions to the aircraft crews. Furthermore, the problem of collision risk modeling in the analysis of "high traffic density" ATC scenarios is different to that of "procedural scenarios", which have been developed by Reich [4] and Brooker [5], amongst others. This is mainly due to the active role of Controllers in the first case. In this case positive control is used extensively to modify the planned aircraft route. This requires

These "collision risk models" were initially applied in the 60s to determine safe separation standards between pairs of aircraft flying at the same altitude on parallel courses over the North Atlantic Ocean [6]. Since then, new models have been developed and continually refined and improved. They have been applied for different geographic regions (USA [7], European airspace [8]), for oceanic or radar [9] environments, and different flight regimes

**Encounters in High Density Airspaces** 

R. Arnaldo, F.J. Sáez, E. Garcia and Y. Portillo

Additional information is available at the end of the chapter

flying nearby in order to estimate the risk of collision.

the inclusion in the model of "human factor response" behavior.

http://dx.doi.org/10.5772/48685

**1. Introduction** 

this field.


http://www.springerlink.com/content/g6122786tjq7w81t/fulltext.pdf
