**2. The treadmill controlled by a patient's walk**

The American Thoracic Society (ATS) in the report issued in 2002, appreciates the advantages of the 6MWT on a treadmill as it saves space and allows constant monitoring during the exercise (ATS, 2002). However, ATS has not approved the use of a treadmill to determine the six-minute walking distance (6MWD) because so far, patients have been unable to pace themselves on an ordinary treadmill.

The popularity of the 6MWT in clinical practice (Stevens et al., 1999; Montogomery & Gardnem, 1998; Roul et al., 1998; Zugck et al., 2000; Rostagno et al.,2003), problems with the performance of the test on the treadmill, as well as the differences between the 6MWT in the hallway and on the treadmill encouraged us to develop a treadmill which applies the algorithm of the safe speed adjustment to the walking capacity of the patient. The purpose of our work was to construct a treadmill which enables the patients to move at their own pace during the walk, as well as to check if such a treadmill would be sufficient to perform the 6MWT.

The paper describes the reasons and a series of works which have led to the development of a new treadmill type adapted to the performance of the 6MWT. The new treadmill, which allows the patients to walk at their own pace, could be useful in rehabilitation, evaluation of physical efficiency, sport training and recreation.

#### **2.1 Preliminary works**

The common treadmill forced patients to adjust their walking speed to its belt speed so as to prevent them from being pushed off the belt. The new treadmill changes its belt speed with

treadmill belt to the pace of the patient's walk by means of the appropriate program and sensors. In this way, a situation equivalent to a hallway walk will be reconstructed where the patient slows down or even stops if tired. The team of constructors from ITAM (Institute of Medical Technology and Equipment) in Zabrze started research into this idea. As a result of their work, a treadmill was constructed which adapts to the walking capacity of a patient suffering from chronic obstructive pulmonary disease (COPD), heart failure (HF) or arterial circulation failure in lower limbs. The algorithm designed to control the speed of the treadmill

The second part of the chapter demonstrates the results of the engineers' work on the construction of the treadmill in order to perform the walk test safely. The third part of the chapter contains the evaluation of adjustment of the treadmill to the walking pace of healthy volunteers, as well as a comparison of the distance covered during the 6-minute walk on the

The obtained results demonstrating the advantages of the treadmill, have encouraged us to perform a 6MWT for patients with heart failure in the II-III NYHA classes (unpublished trial). The treadmill test was tolerated equally well by the patients as the hallway test. The fact that a similar distance was covered in both tests demonstrates that the technological barrier preventing us from obtaining credible results of the six-minute walk test on the treadmill, has been overcome. Thus, the possibility of performing 6MWT on the treadmill for patients suffering from obstructive airways disease, heart failure, or intermittent

The American Thoracic Society (ATS) in the report issued in 2002, appreciates the advantages of the 6MWT on a treadmill as it saves space and allows constant monitoring during the exercise (ATS, 2002). However, ATS has not approved the use of a treadmill to determine the six-minute walking distance (6MWD) because so far, patients have been

The popularity of the 6MWT in clinical practice (Stevens et al., 1999; Montogomery & Gardnem, 1998; Roul et al., 1998; Zugck et al., 2000; Rostagno et al.,2003), problems with the performance of the test on the treadmill, as well as the differences between the 6MWT in the hallway and on the treadmill encouraged us to develop a treadmill which applies the algorithm of the safe speed adjustment to the walking capacity of the patient. The purpose of our work was to construct a treadmill which enables the patients to move at their own pace during the walk, as

The paper describes the reasons and a series of works which have led to the development of a new treadmill type adapted to the performance of the 6MWT. The new treadmill, which allows the patients to walk at their own pace, could be useful in rehabilitation, evaluation of

The common treadmill forced patients to adjust their walking speed to its belt speed so as to prevent them from being pushed off the belt. The new treadmill changes its belt speed with

well as to check if such a treadmill would be sufficient to perform the 6MWT.

belt is based on precise, wireless determination of the patient's position on the belt.

treadmill and in the hallway.

claudication, has appeared.

**2. The treadmill controlled by a patient's walk** 

unable to pace themselves on an ordinary treadmill.

physical efficiency, sport training and recreation.

**2.1 Preliminary works** 

the patient's changing walking speed. It is done quickly enough to keep the patient still on the treadmill. At the beginning, the preparation of such a treadmill seemed difficult for us and a different solution was chosen. We decided to combine the speed of the belt with the patient's position on the treadmill. When the patient is close to the front of the treadmill, the maximum speed of the belt is achieved, when he is close to the end of treadmill, the belt stops. While increasing the walking speed, the patient moves toward the front of the treadmill and the belt speed increases, when he slows down, the belt moves him backwards and the belt speed adjusts again. In order to realize such an algorithm, precise measurements of the patient's position on the treadmill are necessary. Already at the beginning, a decision was made to measure the position wirelessly because all the other methods were rather inconvenient.

First of all, the use of the ultrasound wave, reflected by the patient on the treadmill, was chosen to measure the distance from the patient to the front of the treadmill. Due to parasitical echoes from the other objects around, measurements were uncertain and the method turned out to be inconvenient. Because transmitting and receiving ultrasonic waves is simple and cheap in practice, we decided to continue using ultrasound after some modification. In the new method, the patient was carrying a transmitter which produced simultaneously a short impulse of ultrasound wave and infrared beam (about 100 milliseconds long). Both signals were received by the receiver at the front of the treadmill and the distance between the transmitter and receiver was calculated from the time delay between the received signals. Distance measurements turned out to be accurate (error less than 10 mm) and due to the shortest direct way of the ultrasound signal, parasitical echoes did not interfere with the measurements.

Fig. 1 shows the idea of the patient's position measurement using a mixed ultrasound/ infrared method. Carrying a transmitter seemed slightly uncomfortable for the patient, but

Fig. 1. Six-minute walk test on the treadmill

The Six-Minute Walk Test on the Treadmill 221

patient's stable position on the treadmill regardless of the walking speed by means of a controller utilizing a PID algorithm (proportional – integral – derivative). The algorithm takes into consideration the maximum allowed belt speed set in the console by the operator and the maximum allowed distance from the receiver in which the belt stops. This distance lets the patient stop without a risk of falling off the treadmill. Fig. 3 shows a block diagram

Several blocks have been added to the standard PID controller structure. First, the insensibility area block prevents the PID's operation when the difference between the

Fig. 2. Block diagram of a treadmill controlled by the patient's walk

of the PID control algorithm.

Fig. 3. PID controller algorithm

safe, because only the person carrying the transmitter could operate the treadmill. Additionally, an infrared beam was used to transmit the patient's heart rate (HR) displayed at the control panel of the treadmill.

As soon as the position measurement system with the transmitter and receiver was ready, we began to prepare the treadmill to control the speed. We adapted the treadmill ERT-100 constructed by ITAM, by connecting the receiver and introducing a new control program to its console. By using "the 6-minute walk test program" in the control panel, the operator could input the maximum speed of the treadmill belt and start the test. After 6 minutes from the start of the test program, the test came to an end and displayed the distance covered by the patient. The treadmill belt achieved its maximum speed when the distance between the receiver and transmitter was less than 30 cm and stopped when the distance was longer than 120 cm. Between those two distances, the speed was changing proportionately from 0 to the maximum value which could not exceed 10 km per hour. The ERT-100 treadmill, modified as described above, was examined by a group of 6 healthy volunteers, the employees of ITAM in Zabrze.

The volunteers carried out the 6MWT on the treadmill and in the hallway on two separate days, according to the protocol described in (Lipkin et al., 1986). Distances covered by the volunteers in both tests were similar and the participants found the treadmill test more comfortable than the hallway test. Although the results were satisfactory (Redelmeier et al., 1997), some drawbacks of the treadmill and control algorithm appeared:


It was obvious that we could not change the treadmill to eliminate all of these disadvantages, so we had to prepare a new control algorithm instead.

### **2.2 Main results**

A new idea of treadmill control is shown in Fig.2 as a block diagram. The distance measurement method has remained the same, however, the receiver has got some additional functions. As opposed to the former version, the console is used only for communication with the operator.

The treadmill control algorithm is realized by a microcontroller in the receiver, while the console transfers only the speed and slope signal to the treadmill.

The belt speed change ramp, previously programmed in the console, can now be changed as required for the test. We decided to return to the idea of a treadmill that changes its belt speed along with the patient's changing walking speed – one that does it quickly enough to keep the patient still on the treadmill. The microcontroller in the receiver maintains the

safe, because only the person carrying the transmitter could operate the treadmill. Additionally, an infrared beam was used to transmit the patient's heart rate (HR) displayed

As soon as the position measurement system with the transmitter and receiver was ready, we began to prepare the treadmill to control the speed. We adapted the treadmill ERT-100 constructed by ITAM, by connecting the receiver and introducing a new control program to its console. By using "the 6-minute walk test program" in the control panel, the operator could input the maximum speed of the treadmill belt and start the test. After 6 minutes from the start of the test program, the test came to an end and displayed the distance covered by the patient. The treadmill belt achieved its maximum speed when the distance between the receiver and transmitter was less than 30 cm and stopped when the distance was longer than 120 cm. Between those two distances, the speed was changing proportionately from 0 to the maximum value which could not exceed 10 km per hour. The ERT-100 treadmill, modified as described above, was examined by a group of 6 healthy volunteers, the

The volunteers carried out the 6MWT on the treadmill and in the hallway on two separate days, according to the protocol described in (Lipkin et al., 1986). Distances covered by the volunteers in both tests were similar and the participants found the treadmill test more comfortable than the hallway test. Although the results were satisfactory (Redelmeier et al.,

 the treadmill ERT-100 was too short for such a control algorithm; an additional sloping platform was necessary for the patient's safety, the patient's rapid stop could be

the range of position change was narrow, which caused restless movement of the belt;

the belt speed change ramp was too slow and the treadmill reacted too slowly to rapid

the illusion of similarity to a hallway walk was partial, the patient always had to pay

It was obvious that we could not change the treadmill to eliminate all of these

A new idea of treadmill control is shown in Fig.2 as a block diagram. The distance measurement method has remained the same, however, the receiver has got some additional functions. As opposed to the former version, the console is used only for

The treadmill control algorithm is realized by a microcontroller in the receiver, while the

The belt speed change ramp, previously programmed in the console, can now be changed as required for the test. We decided to return to the idea of a treadmill that changes its belt speed along with the patient's changing walking speed – one that does it quickly enough to keep the patient still on the treadmill. The microcontroller in the receiver maintains the

1997), some drawbacks of the treadmill and control algorithm appeared:

at the same time, slight changes caused perceptible change of speed;

disadvantages, so we had to prepare a new control algorithm instead.

console transfers only the speed and slope signal to the treadmill.

at the control panel of the treadmill.

employees of ITAM in Zabrze.

dangerous for him;

**2.2 Main results** 

changes of the patient's speed;

communication with the operator.

attention to maintaining the desired speed.

Fig. 2. Block diagram of a treadmill controlled by the patient's walk

patient's stable position on the treadmill regardless of the walking speed by means of a controller utilizing a PID algorithm (proportional – integral – derivative). The algorithm takes into consideration the maximum allowed belt speed set in the console by the operator and the maximum allowed distance from the receiver in which the belt stops. This distance lets the patient stop without a risk of falling off the treadmill. Fig. 3 shows a block diagram of the PID control algorithm.

Fig. 3. PID controller algorithm

Several blocks have been added to the standard PID controller structure. First, the insensibility area block prevents the PID's operation when the difference between the

The Six-Minute Walk Test on the Treadmill 223

Probably the treadmills should be equipped with several algorithms, each of them focused on a different group of persons (elderly or young, fit or unfit). A treadmill with a CE certificate allows us to choose the best control algorithm for each group of people through

The "6-minute walk test" is not the only application of the treadmill controlled by the patient's walk. It is easy to imagine many other tests based on the same principle, for example a "2-kilometer walk test" where the principle of 6MWT is inverted and the time elapsed after covering the distance is the final result. By adding elevation changes of the treadmill to the walk, we obtain a system which simulates a cross country test, unknown in

ITAM is currently working on a new family of treadmills with new features and enhanced characteristics (e.g. enhanced dimensions, higher speed). The new treadmills will be equipped with an algorithm which controls the belt speed not only while the patient is walking, but also while he is running. Position measurements are essential if the treadmill is controlled by the patient's walk. Works are in progress on a new family of treadmills using other methods to determine the patient's position. The achieved results seem to be the same

Fig. 4. LabView simulation of treadmill belt speed control

research carried out on a large population.

medicine but well known in sport training.

or even more promising than before.

patient's position and the preset position is negligible. Second, the speed limit block prevents the patient from achieving belt speeds that could be dangerous. The limit is preset by the treadmill operator. Third, the position limit block stops the treadmill belt when the patient's position is outside the controlling range. The same happens if the transmitter signal is lost, e.g. when the transmitter is out of range or inoperative. Another block which prevents the integrator from saturation, stops integration when the belt speed is out of range. The last block resets the controller's integral and derivative blocks.

The above diagram uses the following symbols:


The equations which describe each part of the PID controller are as follows:

$$\text{P:} \qquad \quad \quad \quad \quad \quad \quad \quad \quad \mathbf{U}p(\mathbf{s}) \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad U(\mathbf{s}) = Up(\mathbf{s}) + U\mathbf{i}(\mathbf{s}) + U\mathbf{d}(\mathbf{s})$$

$$\text{I:}\\
\qquad Ui(\mathbf{s}) = k\_p \cdot \frac{1}{\mathbf{s} \cdot \mathbf{T}\_i} \cdot \mathbf{e}(\mathbf{s})\\
\qquad \qquad \qquad \qquad \qquad Ud(\mathbf{s}) = k\_p \cdot \frac{\mathbf{s} \cdot \mathbf{T}\_r}{1 + \mathbf{s} \cdot \mathbf{T}\_d} \cdot \mathbf{e}(\mathbf{s}).$$

The treadmill control using the PID controller described above has been simulated with the use of LabView software. The simulation enables us to specify rough parameters of the controller. Fig. 4 shows the LabView screen during a simulation of the treadmill belt speed control. A program for the PID controller was implemented, with constants specified during simulation, into the receiver's microcontroller and checked by volunteers walking on the treadmill at varying speeds. During the tests, the PID controller parameter was being tuned to achieve the best results.

The personnel performing the tests often had different subjective views on the best set of controller constants. Each person had his own favourite algorithm. The differences seemed to be negligible at the time, however, they will become more meaningful if there are differences between patients (due to age, incapacity, disability).

At the end of the test round, an additional test was carried out, using the treadmill's ability to change its elevation. During the 6MWT, after covering part of the route preset in the program, the treadmill changed its elevation in accordance with the value specified in the program. The test showed that elevation changes did not affect the performance of the test.

As a result, we have developed an algorithm that makes the 6-minute walk test on the treadmill much safer than at the beginning and very similar to a classic hallway test. It has allowed us to prepare a commercial version of the ERT-100 treadmill equipped with a transmitter and receiver. The treadmill has passed the CE certification procedure.

patient's position and the preset position is negligible. Second, the speed limit block prevents the patient from achieving belt speeds that could be dangerous. The limit is preset by the treadmill operator. Third, the position limit block stops the treadmill belt when the patient's position is outside the controlling range. The same happens if the transmitter signal is lost, e.g. when the transmitter is out of range or inoperative. Another block which prevents the integrator from saturation, stops integration when the belt speed is out of

range. The last block resets the controller's integral and derivative blocks.

The equations which describe each part of the PID controller are as follows:

**E –** error, difference between the preset position and current distance

The treadmill control using the PID controller described above has been simulated with the use of LabView software. The simulation enables us to specify rough parameters of the controller. Fig. 4 shows the LabView screen during a simulation of the treadmill belt speed control. A program for the PID controller was implemented, with constants specified during simulation, into the receiver's microcontroller and checked by volunteers walking on the treadmill at varying speeds. During the tests, the PID controller parameter was being tuned

*sU Up s Ui s Ud s*)()()()(

)( <sup>1</sup> )( *se Ts*

*d r <sup>p</sup>*

*Ts ksUd*

The personnel performing the tests often had different subjective views on the best set of controller constants. Each person had his own favourite algorithm. The differences seemed to be negligible at the time, however, they will become more meaningful if there are

At the end of the test round, an additional test was carried out, using the treadmill's ability to change its elevation. During the 6MWT, after covering part of the route preset in the program, the treadmill changed its elevation in accordance with the value specified in the program. The test showed that elevation changes did not affect the performance of the test. As a result, we have developed an algorithm that makes the 6-minute walk test on the treadmill much safer than at the beginning and very similar to a classic hallway test. It has allowed us to prepare a commercial version of the ERT-100 treadmill equipped with a

transmitter and receiver. The treadmill has passed the CE certification procedure.

**Y** – the distance between the transmitter and receiver **Sp –** the preset patient's position on the treadmill

The above diagram uses the following symbols:

**Kp –** the controller proportional gain **TI –** the controller integral time **Tr –** the controller derivative time

P: D:

I: U:

*i*

)( <sup>1</sup> )( *se Ts*

*<sup>p</sup>*

differences between patients (due to age, incapacity, disability).

**U –** the treadmill belt speed

**Td –** the inertia time base

*Up seks* )()( *<sup>p</sup>*

*Ui ks*

to achieve the best results.

Fig. 4. LabView simulation of treadmill belt speed control

Probably the treadmills should be equipped with several algorithms, each of them focused on a different group of persons (elderly or young, fit or unfit). A treadmill with a CE certificate allows us to choose the best control algorithm for each group of people through research carried out on a large population.

The "6-minute walk test" is not the only application of the treadmill controlled by the patient's walk. It is easy to imagine many other tests based on the same principle, for example a "2-kilometer walk test" where the principle of 6MWT is inverted and the time elapsed after covering the distance is the final result. By adding elevation changes of the treadmill to the walk, we obtain a system which simulates a cross country test, unknown in medicine but well known in sport training.

ITAM is currently working on a new family of treadmills with new features and enhanced characteristics (e.g. enhanced dimensions, higher speed). The new treadmills will be equipped with an algorithm which controls the belt speed not only while the patient is walking, but also while he is running. Position measurements are essential if the treadmill is controlled by the patient's walk. Works are in progress on a new family of treadmills using other methods to determine the patient's position. The achieved results seem to be the same or even more promising than before.

The Six-Minute Walk Test on the Treadmill 225

in a horizontal position and the belt speed was controlled by constant measurement of the

After 6 minutes from the start of the test, the program ended the test and displayed the

The test on the treadmill was preceded by a training session lasting a few minutes on the day before the actual test. During the training session the participant learned how the treadmill worked and walked a distance of 100 meters at a changeable pace, as well as

The participants were informed about the treadmill test in an identical way as about the

The comfort of the test and the distance covered in metres were subject to evaluation in both cases. The evaluation scale for comfort included the question of which type of test was less problematic during performance or whether the comfort of both tests was so similar that the differences were negligent. The number of indications to a given type of test was calculated. The treadmill was also monitored from the point of view of smooth speed adjustment to the

The pulse and blood pressure were measured before and after each test in order to assess

The aim of the statistical analysis was to compare the values of the distance covered, obtained in both 6MWT varieties. Also the heart rate and blood pressure before and after the

Multidimensional statistical research was conducted, as well the T2 test was applied for vectors of the expected values for both varieties in order to verify whether the compared

The comfort of the treadmill test was indicated as better by 18/29 of the participants, the hallway test was indicated as better by 4/29 of the participants and both tests were

During the test, healthy volunteers were walking most frequently with the speed of 7 km/hour (4÷10). The average distance covered on the treadmill was 683**.**0 m and was usually 57**.**1 m longer on average than in the hallway (Table 1). This difference turned out to be statistically significant. The participants covered 29 laps on average during the hallway

No considerable difference could be seen in the heart rate before the tests. Also, the resulting accelerated heart rate after both types of tests did not show any marked difference (Table 2),

test were compared using the Student's test for matched pairs for independent trials.

individual's sudden slowdown without affecting his or her balance.

evaluated as identical in terms of comfort by 7/29 of the participants.

the hemodynamic impact of both 6MWT varieties.

research leads to similar hemodynamic consequences.

patient's position on the treadmill.

distance covered by the patient.

**3.1.4 Analyzed parameters** 

**3.1.5 Statistical analysis** 

**3.2 Results** 

test (19÷36).

just like blood pressure (Table 2).

hallway test.

practiced stopping and restarting the walk.
