**5.1 Experiment by simulator**

Examination environment for the Ripple detection method and the Stationary net detection method is shown in Fig. 6 and Fig. 7. There are twelve camera nodes in the environment of floor map 1, and there are fourteen camera nodes in the environment of floor map 2. Here, the following conditions are set in order to examine the effectiveness of these detection methods. i) Camera nodes are arranged on latticed floor, 56*m ×* 56*m*. ii) View distance of camera is set to 10*m* in one direction. iii) Identification of a target in the image processing does not fail when re-detecting. iv) Walking speed of the target is constant. v) Only one target is searched. vi) The target moves only forward without going back. In the case of the floor map 1, the target moves following the order of *a*1, *a*2, *a*3, *a*4, *a*5, *a*6, *a*7, *a*8, *a*9, *a*10, *a*11, *a*12 and *a*1. In the case of the floor map 2, the target moves following the order of *a*1, *a*2, *a*4, *a*5, *a*7, *a*9, *a*10, *a*11 and *a*1. In the examination, the time that an agent concludes a failure of tracking is same as search cycle time. The search cycle time is defined as the time concluded that an agent can not discover a target. The search cycle time is prepared using 3 patterns 12 seconds, 9 seconds and 6 seconds. Walking speed of the target is prepared using 3 patterns 1*.*5*m/s*, 2*m/s* and 3*m/s*. And search of target is prepared that an agent loses a target at *a*7 and the agent starts a search in the situation that the target has already moved to *a*8. Furthermore, Stationary net detection method is examined by 3 patterns *n* = 2, *n* = 3 and *n* = 4, because of confirming effectiveness by number of non-camera nodes. On each floor map, using 12 patterns of such combination by each walking speed, discovery time and the number of agents are measured. Generally, the walking speed of a person is around 2*.*5*m/s*, and the two types of walking speed, 2*m/s* and 3*m/s*, used by the target which was examined are almost equivalent to the walking speed of general person. And walking speed, 1*.*5*m/s*, is very slow from the walking speed of general person.

The results of the measurement on the floor map 1 are shown in Table 1, Table 2 and Table 3. The results of the measurement on the floor map 2 are shown in Table 4, Table 5 and Table 6. They are a mean value of 5 measurements.

The result of the Ripple detection method shows that the discovery time becomes shorter and usual tracking can resume more quickly, if the target exists near where the agent lost. But, if the walking speed of a target is faster, the agent will become difficult to discover the target.

The result of the Stationary net detection method shows that the agent can discover a target if coefficient *n* has larger value, even if the walking speed of a target is faster. And it is not enough interval to re-detect a target if *n* consists of *n ≤* 3 and it is not enough time to redetect the target if the search cycle time is shorter.

From the result of measurement on the floor map 1, if the Stationary net detection method uses coefficient *n* = 4, there is not the difference of efficiency between the Ripple detection method and the Stationary net detection method. However, from the result of measurement

Fig. 6. Floor map 1 for experiment of detection methods.

30 Recent Developments in Video Surveillance

detection methods are experimented, and the effectiveness is verified. In the experiment by real environment, the tracking method is verified for whether the plural targets can be

Examination environment for the Ripple detection method and the Stationary net detection method is shown in Fig. 6 and Fig. 7. There are twelve camera nodes in the environment of floor map 1, and there are fourteen camera nodes in the environment of floor map 2. Here, the following conditions are set in order to examine the effectiveness of these detection methods. i) Camera nodes are arranged on latticed floor, 56*m ×* 56*m*. ii) View distance of camera is set to 10*m* in one direction. iii) Identification of a target in the image processing does not fail when re-detecting. iv) Walking speed of the target is constant. v) Only one target is searched. vi) The target moves only forward without going back. In the case of the floor map 1, the target moves following the order of *a*1, *a*2, *a*3, *a*4, *a*5, *a*6, *a*7, *a*8, *a*9, *a*10, *a*11, *a*12 and *a*1. In the case of the floor map 2, the target moves following the order of *a*1, *a*2, *a*4, *a*5, *a*7, *a*9, *a*10, *a*11 and *a*1. In the examination, the time that an agent concludes a failure of tracking is same as search cycle time. The search cycle time is defined as the time concluded that an agent can not discover a target. The search cycle time is prepared using 3 patterns 12 seconds, 9 seconds and 6 seconds. Walking speed of the target is prepared using 3 patterns 1*.*5*m/s*, 2*m/s* and 3*m/s*. And search of target is prepared that an agent loses a target at *a*7 and the agent starts a search in the situation that the target has already moved to *a*8. Furthermore, Stationary net detection method is examined by 3 patterns *n* = 2, *n* = 3 and *n* = 4, because of confirming effectiveness by number of non-camera nodes. On each floor map, using 12 patterns of such combination by each walking speed, discovery time and the number of agents are measured. Generally, the walking speed of a person is around 2*.*5*m/s*, and the two types of walking speed, 2*m/s* and 3*m/s*, used by the target which was examined are almost equivalent to the walking speed of general person. And walking speed, 1*.*5*m/s*, is

The results of the measurement on the floor map 1 are shown in Table 1, Table 2 and Table 3. The results of the measurement on the floor map 2 are shown in Table 4, Table 5 and

The result of the Ripple detection method shows that the discovery time becomes shorter and usual tracking can resume more quickly, if the target exists near where the agent lost. But, if the walking speed of a target is faster, the agent will become difficult to discover the

The result of the Stationary net detection method shows that the agent can discover a target if coefficient *n* has larger value, even if the walking speed of a target is faster. And it is not enough interval to re-detect a target if *n* consists of *n ≤* 3 and it is not enough time to re-

From the result of measurement on the floor map 1, if the Stationary net detection method uses coefficient *n* = 4, there is not the difference of efficiency between the Ripple detection method and the Stationary net detection method. However, from the result of measurement

tracked continuously.

**5.1 Experiment by simulator** 

very slow from the walking speed of general person.

Table 6. They are a mean value of 5 measurements.

detect the target if the search cycle time is shorter.

target.

on the floor map 2, if a floor map is complicated, the discovery time of the Stationary net detection method becomes shorter than the discovery time of the Ripple detection method and the number of agents of the Stationary net detection method becomes less than the number of agents of the Ripple detection method.

On the whole, the result of both methods shows that a number of agents decreases by searching a target near search cycle but the agents can not search the target if the search cycle time is longer than the waking speed. In addition based on the results, when the walking speed is faster, the discovery time is shortened or equal and the number of agents decreases or is equal.


Table 1. Detection time on floor map 1 by walking speed 1.5*m/s.* 

A Construction Method for Automatic Human Tracking System with Mobile Agent Technology 33

Search Cycle(12*s*) Number of Agents 14 14 10 12.2

Search Cycle(9*s*) Number of Agents 13.8 10 10 13

Search Cycle(6*s*) Number of Agents 13.9 10 13.2 14

Search Cycle(12*s*) Number of Agents 14 14 10 10

Search Cycle(9*s*) Number of Agents 14 14 10 13

Search Cycle(6*s*) Number of Agents 13.8 10 13 14

Search Cycle(12*s*) Number of Agents 14 14 14 9.8

Search Cycle(9*s*) Number of Agents 14 14 14 10

Search Cycle(6*s*) Number of Agents 14 14 10.4 12.4

Upon verification of real system, it aimed to confirm whether targets can be tracked continuously by tracking method. Therefore, in order to reduce influence by image processing performance, the image processing adopts simple processing by color vision. In this experiment, there is a miniature environment of 2*m* by 2*m* shown in Fig. 8 and Fig. 9. Here are three targets and those targets have each moving routes, red, blue, and green, shown in Fig. 8. The environment consists of seven servers with USB camera. The toys are used as targets instead of people tracked. This toy is a train toy with a sensor that recognizes a black line. And the train runs on along the line. The black line is used as a route a target walks and is drawn by hand. The toy is covered with a color paper to keep a certain accuracy of image processing as shown in Fig. 9. The moving speed of target is 6.8*m/second*. This environment has simulated a floor of 60*m* by 60*m* on the scale of 1/30; therefore the

Net (n=2)

Net (n=2)

Net (n=2)

Discovery Time (*s*) 49.5 - 31.8 32.6

Discovery Time (*s*) 33.7 32.3 32.3 33

Discovery Time (*s*) 32.2 32 32.4 33

Discovery Time (*s*) - - 31.1 25.3

Discovery Time (*s*) 35.9 - 25.6 25.2

Discovery Time (*s*) 24.8 25.3 24.8 25.8

Discovery Time (*s*) - - - 18

Discovery Time (*s*) - - - 18.2

Discovery Time (*s*) 25.9 - 18.2 19

Stationary Net (n=3)

Stationary Net (n=3)

Stationary Net (n=3) Stationary Net (n=4)

Stationary Net (n=4)

Stationary Net (n=4)

Walking Speed (1.5*m/s*) Ripple Stationary

Walking Speed (2m/s) Ripple Stationary

Walking Speed (3*m/s*) Ripple Stationary

Table 4. Detection time on floor map 2 by walking speed 1.5*m/s.* 

Table 5. Detection time on floor map 2 by walking speed 2*m/s.* 

Table 6. Detection time on floor map 2 by walking speed 3*m/s* 

**5.2 Experiment by real environment** 

Fig. 7. Floor map 2 for experiment of detection methods.


Table 2. Detection time on floor map 1 by walking speed 2*m/s.* 


Table 3. Detection time on floor map 1 by walking speed 3*m/s.* 


Table 4. Detection time on floor map 2 by walking speed 1.5*m/s.* 

32 Recent Developments in Video Surveillance

Fig. 7. Floor map 2 for experiment of detection methods.

Table 2. Detection time on floor map 1 by walking speed 2*m/s.* 

Table 3. Detection time on floor map 1 by walking speed 3*m/s.* 

Walking Speed (2*m/s*) Ripple Stationary

Walking Speed (3*m/s*) Ripple Stationary

Search Cycle(12*s*) Number of Agents 6 12 12 6

Search Cycle(9*s*) Number of Agents 6 12 12 6

Search Cycle(6*s*) Number of Agents 7 12 6 7

Search Cycle(12*s*) Number of Agents 12 12 12 12

Search Cycle (9*s*) Number of Agents 5.9 12 12 6

Search Cycle(6*s*) Number of Agents 6 12 12 6

Net (n=2)

Net (n=2)

Discovery Time (*s*) 15.5 - - 15.5

Discovery Time (*s*) 15.5 - - 15.4

Discovery Time (*s*) 16.5 - 15.5 15.7

Discovery Time (*s*) - - - -

Discovery Time (*s*) 11.7 - - 11.5

Discovery Time (*s*) 11.7 - - 11.6

Stationary Net (n=3)

Stationary Net (n=3) Stationary Net (n=4)

Stationary Net (n=4)


Table 5. Detection time on floor map 2 by walking speed 2*m/s.* 


Table 6. Detection time on floor map 2 by walking speed 3*m/s* 
