Wireless Biomedical Signal Monitoring Device on Wheelchair ... - IPS

Proceedings of the 29th Annual International Conference of the IEEE EMBS Cité Internationale, Lyon, France August 23-26, 2007.

ThP2A1.11

Wireless Biomedical Signal Monitoring Device on Wheelchair using Noncontact Electro-mechanical Film Sensor Jong-Myoung Kim, Joo-Hyun Hong, Myeong-Chan Cho, Eun-Jong Cha, Tae-Soo Lee, Member, IEEE Abstract—The present study purposed to measure the BCG (Ballistocardiogram) of subjects on a wheelchair using a noncontact electro-mechanical film sensor (EMFi sensor) and detect the respiratory rate from BCG in real-time while the subjects are moving. In order to measure wirelessly the BCG of subjects moving on a wheelchair, we made a seat-type noncontact EMFi sensor and developed a transmitter and a receiver using Zigbee wireless RF communication technology. The sensor is embedded with a 3-axis accelerometer to remove the noise of wheelchair vibration from BCG signal. Signal obtained from each sensor goes through the A/D converter and is recorded in the SD (Secure Digital) card in PDA (Personal Digital Assistance) with a receiving part. We also developed a PC (Personal Computer) data analysis program, analyzed data recorded in the SD card using the program, and presented the results in graph. Lastly, this study demonstrated that a warning message can be sent from PDA to the remote server via a CDMA (Code Division Multiple Access) network in case the person on wheelchair falls in emergency. Our experiment was carried out with healthy male and female adults in their 20s who volunteered to help this research. The results of analyzing collected data will show that the respiratory rate can be measured in real-time on a moving wheelchair.

developed equipment, stored the data in the SD card of PDA connected to the receiving part, and analyzed the data. The experiment was carried out with 4 healthy male and female adults in their 20s who volunteered to help the research. The volunteers were seated on a wheelchair, and while they kept still signals from SKT (skin temperature amplifier) sensor provided by Biopac Company were compared with signals from the equipment developed by our research team. The objective of the comparative experiment was to demonstrate the appropriateness of the developed equipment. After comparison with respiratory rate and heart rate data obtained from the SKT sensor, our experiment also demonstrated that the respiratory rate can be measured on a moving wheelchair in three different environments. Lastly, this study described the process that a warning message is delivered from the PDA phone to the remote server via a CDMA network. Ultimately, this study aimed to demonstrate that the respiratory rate of physically vulnerable wheelchair users including the aged and the weak can be measured in real-time while they are moving and based on the information proper actions can be taken promptly when such patients fall in emergency.

I. INTRODUCTION

II. HARDWARE

BCG (ballistocardiogram) is a noncontact cardiac cycle measuring method without using electrode [1]. In general, BCG is composed of cardiac impulse signal, respiratory signal and kinetic signal [2, 3]. However, because the main objective of the present study is to show that the respiratory rate can be measured in real-time on a moving wheelchair, we detected only respiratory signal, excluding kinetic and cardiac impulse signals contained in BCG signal. To obtain BCG signal, we converted data from the transmission part of This study was supported by a grant from the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea (Grant No A040032). Jong-Myoung Kim is with Department of Biomedical Engineering College of Medicine, Chungbuk National University, Cheongju, South Korea (e-mail: [email protected]) Joo-Hyun Hong is with Department of Biomedical Engineering, College of Medicine, Chungbuk National University, Cheongju, South Korea (e-mail: [email protected]). Myeong-Chan Cho is with Department of Internal Medicine, College of Medicine, Chungbuk National University, Cheongju, South Korea (e-mail: [email protected]). Eun-Jong Cha is with Department of Biomedical Engineering, College of Medicine, Chungbuk National University, Cheongju, South Korea (e-mail: [email protected]). Tae-Soo Lee is with Department of Biomedical Engineering, College of Medicine, Chungbuk National University, Cheongju, South Korea (phone: +82-43-269-6332; fax: +82-43-272-6332; e-mail: tslee@ chungbuk.ac.kr).

1-4244-0788-5/07/$20.00 ©2007 IEEE

A. Hardware 1) Making a seat-type noncontact EMFi sensor for a wheelchair In this study, we made a seat-type EMFi sensor and transmitter that can be installed on a wheelchair so that the patient’s BCG can be measured without the patient’s contact. The seat is laid on the cushion of the wheelchair on which the patient is seated, and measures the patient’s BCG. 2) Transmission part The hardware part was composed of the transmission part and the receiving part. The transmission part is again composed of microcontroller, Zigbee wireless RF communication chip, and EMFi sensor. In addition, it includes an accelerometer to remove noise occurring from the movement of the wheelchair. The EMFi sensor is made of polypropylene, and the inside of the film has cell-shaped structure. The sensor uses the mechanism that if pressure is applied from outside the thickness of air-void changes. The physical mechanism of the sensor is as follows. ∆q = k ∆F

(1)

In Equation (1), ∆q is the generated charge. k is the

574

Authorized licensed use limited to: UNIVERSIDADE TECNICA DE LISBOA. Downloaded on January 27, 2009 at 08:49 from IEEE Xplore. Restrictions apply.

sensitivity variable, and ∆F indicates the force applied from outside [4]. The transmission part is powered by two AAA batteries. Figure 1 shows the transmission part developed by the research team, which was installed on the wheelchair.

stored in the PDA through a serial cable.

Fig. 3. Block diagram for acquiring BCG and acceleration data

Fig. 1. The wireless BCG measurement device mounted on the cushion of wheel chair

C. Data transmission to the remote server We developed a transmission system for sending the biomedical signal of a person on a wheelchair to the remote server in real-time through a CDMA network [6]. The remote server analyzes the received signal and delivers instructions from the doctor or counselor to the patient. The patient’s situation can be recognized promptly based on signal delivered to the remote server and the first aid squad can be sent to the patient’s place in time.

3) Receiving part The receiving part was made using microcontroller and Zigbee wireless RF communication chip. In addition, it includes a CDMA module to trigger an event in the remote server when an emergent situation takes place in the patient [5]. The developed receiving part is linked to PDA and displays or stores signals using the PDA. Figure 2 shows that the receiver is connected to PDA through the serial port.3.

Fig. 4. Data transmission to the remote server through a CDMA network in emergency

III.

Fig. 2. Receiving part hardware

B. Software In order to display and store signal in PDA, we developed a PDA application program. From the receiver, the program obtains signal, which is sent in real-time by the transmitter through A/D conversion, and then it stores the data into the SD card as a text file. We also developed a PC program to analyze signal stored in the SD card. The sampling rate of the developed equipment is around 300 Hz. To extract the respiratory rate and the heart rate from original BCG signal, we processed the signal using a Gaussian filter (window size: 350 samples). Figure 3 shows a block diagram for the process that BCG signal and 3-axis acceleration signal from the transmitter are sent to the receiver and then displayed and

RESULTS

A. Comparison of respiratory signal with existing equipment In order to examine the accuracy of signal generated by the equipment that the research developed, we compared the equipment with Biopac’s SKT (skin temperature amplifier) sensor, which is attached between the nose and the upper lip and detects breath using temperature change resulting from inspiration and expiration. The sampling rate of the Biopac’s equipment was set at 200 Hz. For the experiment, we needed a subject and two testers. Experiment was conducted in the following way. A SKT sensor was attached to the subject and the subject was seated on the wheelchair that we developed. Of the two testers, one executed Biopac program embedded with the SKT sensor and the other executed the PDA program developed in this research. The programs were stopped after a minute and the experiment was finished. Results measured using the two sensors simultaneously are as follows.

575 Authorized licensed use limited to: UNIVERSIDADE TECNICA DE LISBOA. Downloaded on January 27, 2009 at 08:49 from IEEE Xplore. Restrictions apply.

2) Experiment 2. Self-driving hand-operated wheelchair In Experiment 2, the subject on the wheelchair drives the hand-operated wheelchair by himself/herself and checks the respiratory rate while moving the wheelchair. The procedure is the same as that of Experiment 1. Figure 7 is the result of Experiment 2, showing the respiratory signal obtained from the subject.

Fig. 7. 20 respirations measured in Experiment 2 for 50 seconds Fig 5. One minute simultaneous recording of SKT and seat sensor data , (a) is using Biopac SKT sensor for 19 respirations for a minute (sampling rate 200), which is (b) BCG signal (sampling rate 300), and (c) filtered respiratory signal

B. Results in three types of environment 1) Experiment 1. Hand-operated wheelchair pushed by a helper To prove that the respiratory rate can be measured in real-time while the wheelchair is moving, we conducted an experiment as follows. Because there was no standard wireless respiratory rate meter, the subject counted his/her own respiratory rate for a minute. Then, the counted respiratory rate was compared with the respiratory rate measured by the sensor on the wheelchair. In Experiment 1, we assumed that there is a helper who is pushing the wheelchair because the patient does not have strength to operate the wheelchair. Accordingly, the subject checks only his/her own respiratory rate. z

Place: The corridor of the laboratory

z

Distance: 48 m

z

Time: Around one minute

z

Average wheelchair speed: 0.8 m/s

Figure 6 is the result of Experiment 1, showing respiratory signal obtained from the subject.

3) Experiment 3. Electrically-driven wheelchair In Experiment 3, the subject operates the electrically-driven wheelchair while riding the wheelchair and at the same time checks the respiratory rate. The procedure is the same as that of Experiment 1. Figure 8 is the result of Experiment 3, showing the respiratory signal obtained from the subject.

Fig. 8. 15 respirations measured in Experiment 3 for one minute

C. Results of comparison among 4 normal persons As in Table 1, the results of analyzing signals from Biopac’s SKT sensor and from the transmission part developed by the research team show that respiratory signal is almost identical except Subject 4. This proves the accuracy of the transmission part, the receiving part and the PDA program developed by our research team. In addition, when the patient’s hand-operated wheelchair is pushed by a helper as in Experiment 1, we can expect that the respiratory rate measured will almost coincide with the actual respiratory rate. Likewise, we can expect the same when a patient operates an electrically-driven wheelchair as in Experiment 3. In Experiment 2, the actual respiratory rate was different from the measured one. This is probably because of noise made by the shaking of the wheelchair and the subject’s body as the subject drove the wheelchair by himself/herself. Table 1 shows the results of comparing the subjects in different experiment conditions.

Fig. 6. 15 respirations measured in Experiment 1 for one minute


TABLE 1. SUBJECTS’ RESPIRATORY RATE IN DIFFERENT EXPERIMENT

[6]

CONDITIONS

Experiment

Comparison

1. Real raspiration/ Measured value



Biopac

Seat sensor

subject1

20/19

17//not measurable

17/16

14

14

subject2

21/20

20/25

21/21

15

15

subject3

16/16

10/19

14/14

16

16

14/11

22//not measurable

25/18

15

12

subject4

IV.

J. H. Hong, N. J. Kim, E. J. Cha, T. S. Lee, “Zigbee based Photoplethysmography”, Journal of Korea Intellectual Patent Society, vol. 8, no. 3, pp.31-35, Sep. 2006.

DISCUSSION AND CONCLUSIONS

The present study installed a seat-type EMFi sensor on a wheelchair and developed a wireless transmission part and a receiving part connected to PDA, stored data obtained from them in the SD card of PDA, and compared the developed equipment with Biopac’s SKT sensor. In addition, the results were analyzed through three experiments. Our experiment was conducted with healthy male and female adults in their 20s who volunteered to help the research. As shown in the results, satisfactory results were obtained from all the subjects except Subject 4. In Experiment 2, the respiratory rate was not measurable for subject 1 and subject 4 because of the shaking of the body and the vibration of the wheelchair as the subject drove the hand-operated wheelchair by themselves. To remove the noise resulting from the movement of the body or the wheelchair, we installed a 3-axis accelerometer, and we plan to perform a research to solve the problem using an accelerometer. In addition, we are developing a system and programs that include a receiving part embedded with a CDMA module to send data to the server in emergency so that actions can be taken promptly to cope with the emergent situation. In conclusion, one of today’s major social concerns is welfare for the aged and the disabled and, in this sense it is considered highly meaningful to extract biomedical signals non-invasively in real-time from patients on their wheelchair and this suggests the importance of this study. REFERENCES [1] [2]

[3]

[4] [5]

X. Yu and D. Dent, “Neural networks in ballistocardiography(BCG) using FPGAs,” in IEE Colloquium on Software Support and CAD Techniques for FPGAs, pp. 7/1-7/5, April 1994. I. Starr, “Further clinical studies with the ballistocardiograph on abnormal form, on digitalis action, in thyroid disease, and in coronary heart disease,” Transaction of the Association of American Physicians, vol. 59, pp. 180-189, 1946. B. M. Baker Jr., W. R. Scarborough, R. E. Mason, et al., “Coronary artery disease studied by ballistocardiography: a comparison of abnormal ballistocardiograms and electrocardiograms,” Transactions of the American Clinical and Climatological Association, vol. 62, p. 191, 1950. J. Lekkala, M. Paajanen, “EMFi-New electret Material for Sensors and Actuators”, IEEE ISE, pp. 743-746, September 1999. J. H. Hong, N. J. Kim, E. J. Cha, T. S. Lee, “A PDA-based Wireless ECG Monitoring System for u-Healthcare”, Journal of Korean Society of Medical Informatics, vol. 12, no. 2, pp.153-159, June 2006.
