Network Approach for Physiological Parameters ... - CiteSeerX

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

337

Network Approach for Physiological Parameters Measurement Fazlur Rahman, Member, IEEE, Arun Kumar, Gangadharan Nagendra, Member, IEEE, and Gourab Sen Gupta

Abstract—A portable parameter monitoring and analysis system for physiological studies and for assisting patient-centric health care management is developed. The system uses the network approach to acquire the data from sensors and transmit them to a server through wireless propagation means. The system automates the acquisition and monitoring of physiological parameters by continuous display on the monitor screen. Programming is done using industry strength software to display data and trends in real time on a standard PC. The measured data is accurate and lives up to the standards of the industry. Index Terms—Biomedical measurements, data acquisition, intelligent sensors, monitoring, network servers, transducers, wireless LAN.

I. INTRODUCTION

T

HERE IS an ever increasing demand for automated monitoring of physiological parameters. The demand is set to increase manifolds in the not-so-distant future in the realm of home care for the aged and patients in hospitals. It will also be needed to monitor the performance level of athletes, service personnel, and others involved in various physical activities. Real-time, continuous monitoring is required in such cases, as it allows for emergency detection of an abrupt change of a person’s health condition. It also monitors and detects changes that take place over long intervals of time, as in the case of athletes. In particular, an ambulatory system that would allow long-term monitoring of mobile subjects is highly in demand. Numerous physiological, biomechanical, and peripheral parameters have been data logged [1] successfully using ambulatory monitoring [2], including electrocardiograph (ECG) [3], electroencephalography (EEG) [4], oesophageal mobility and pH profile [5], planter pressures [6], perspiration measurement [7], blood pressure [8], and bladder pressures [9], among others. A flexible PC-based physiological monitoring system for animal experiments has also been developed [10]. A challenging issue is to satisfy the varying requirements for the collection of different parameters due to differing characteristics of the signal being measured. A ring sensor to monitor heart rate and oxygen saturation in a totally unobtrusive way has been developed [11]. A wearable garment that functions as a wearable

health monitoring system [12] is developed. It can record heart rate, body temperature, motion, position, and barrier penetration. The paper discusses future possible developments in wearable systems, as well as the resulting transformation of healthcare and positive impact on the quality of life of individuals. A wireless approach to a personal health monitoring system based on wireless body area network of intelligent sensors has been developed [13]. PDA-based devices and, more generally, mobile computing tools relying on IEEE 802.11b (wLAN), Bluetooth (wPAN), and cell phone technology (wMAN) are used to integrate wearable technology into neurorehabilitation and consumer-centered mobile telerehabilitation [14]. Among the widely accepted physiological [15] indices, such as temperature, muscle strain, heart rate, and oxygen saturation signals are a few of the important indicators of health condition. These indices can deliver abundant information about the health condition of a patient or an athlete, provided that they can be monitored continuously over a reasonable duration of time, recorded, and analyzed. Although the physiological parameters are measured and monitored locally, the existing systems are bulky and lack the capability of transmitting the data for processing and diagnosis over the internet to a remote location. Usually, physiological data are provided by independent monitors which, in the absence of real-time collation, preclude quick assessment of all physiological variables [16]. In this paper, we describe a new approach for data collection, monitoring, and understanding physiological status through a network approach for assisting patient-centric health care management. To accomplish this, a portable physiological monitoring [17] system, which is able to record physiological parameters such as height, weight, body temperature, heart rate, and blood oxygen saturation level SpO of a person, has been developed. The system can be easily moved to outdoor locations to record these physiological parameters. The data collected is stored in a central computer, transmitted to a server through a wireless network [18], and this data can be accessed over the internet. Also, the data of a person collected over a period of time can be viewed as a trend chart to report the changes over that duration. II. DESIGN

Manuscript received June 15, 2003; revised May 27, 2004. This work was supported by the School of Electrical and Electronic Engineering, Singapore Polytechnic, Singapore. F. Rahman, A. Kumar, and G. Nagendra are with the School of Electrical and Electronic Engineering, Singapore Polytechnic, Singapore. G. S. Gupta is with the School of Electrical and Electronic Engineering, Singapore Polytechnic, Singapore. He is also with the Institute of Information Sciences and Technology, Massey University, Wellington, New Zealand. Digital Object Identifier 10.1109/TIM.2004.834595

A block diagram of the physiological measuring and monitoring system (PMMS) is shown in Fig. 1. The system measures and monitors the physiological parameters gathered from different type of sensors used for this purpose and are listed in Table I. The details of the sensors used for physiological parameter measurement are discussed next.

0018-9456/$20.00 © 2005 IEEE

338

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

Fig. 1.

Block diagram of physiological measuring and monitoring system.

TABLE I FOUR BASIC SENSORS AND THEIR FUNCTIONS FOR PHYSIOLOGICAL PARAMETER MEASUREMENT

A. Pulse Oximeter Arterial blood haemoglobin oxygen saturation and heart rate are measured noninvasively using a pulse oximeter. Post-processing of sensor data can improve usability, as illustrated by recent improvements in pulse oximetry technology [19]. The principle of pulse oximetry is based on the red and infrared light absorption characteristics of oxygenated and deoxygenated (or reduced oxygen) hemoglobin. Typical body sites for pulse oximetry are the finger, toe, pinna (top), and lobe of the ear. The pulse oximeter uses two light emitting diodes (LEDs) that transmit red and infrared light. Red light is in the 600–750-nm-wavelength range, and infrared light is in the 850–1000-nm-wavelength range of the electromagnetic spectrum. The two LEDs are placed on one side of a finger and a photodetector is placed on the opposite side of the finger to receive both the transmitted red (R) and infrared (IR) light passing through the finger. Oxygenated hemoglobin will absorb infrared light and transmit red light. Whereas, deoxygenated hemoglobin will absorb red light and transmit infrared light. An in-built signal processor calculates the ratio of red to infrared (R/IR) signal and compares it to a data table (made up of empirical formulae) and derives an SpO value. Oxygen saturation is determined and averaged over a few seconds, and the recorded value is continuously updated. The pulse oximeter monitors, continuously and noninvasively, the arterial oxygen SpO saturation and the heart rate. It also visualizes their values, the trend in time, and the plethysmograph. It transmits the data via a serial cable to the PC. The acquired data is analyzed by tailor-made software running on the personal computer.

Fig. 2. Timing diagram for proximity detector.

B. Proximity Detector The ultrasound transducer detects the height of the human while standing on the weighing platform. The transducer is mounted on a 2.5-m-long stand pointing toward the weighing scale placed on the floor. It transmits a burst of sound pulses which has a frequency outside the range of human hearing. The pulse travels at the speed of sound (roughly 0.9 ft/s) away from the transducer in a cone shape and the sound reflects back to the detector from any object in the path of this sonic wave. The detector pauses for a brief interval after the sound burst is transmitted and then awaits the reflected sound in the form of an echo. The controller driving the detector initiates a measurement; the detector creates the sound pulse and waits for the return echo. If received, the detector reports this echo to the controller and the controller then computes the distance to the object based on the elapsed time. Basic Timing Diagram The input should be held low (logic 0) and then brought to high for a minimum of 10 s to initiate the sonic pulse. The pulse is generated on the falling edge of this input trigger (Fig. 2). The detector’s receiver circuitry is held in a short blanking interval of 100 s to avoid noise from the initial signal, and then it is enabled to listen for the echo. The falling edge of the echo line signals either an echo detection or a timeout (if no object echo is detected). The controller begins timing on the falling edge of the trigger input and ends timing on the falling edge of the echo line. This duration determines the distance to the first object from which the echo is received. After the transducer is mounted on the stand, the distance between the weighing scale and the transducer is measured by a measuring tape. A 2-m pole is placed on top of the weighing scale, and the distance between the pole and the transducer is measured by a measuring tape. These readings are used to calibrate the transducer by the software that is described in the software section of this paper. C. Temperature Transducer A noncontact infrared probe (Fluke 80T-IR) is used to monitor the tympanic temperature [20]. An infrared thermometer measures radiant energy beyond the sensitive range of the

RAHMAN et al.: NETWORK APPROACH FOR PHYSIOLOGICAL PARAMETERS MEASUREMENT

human eye. All objects radiate this energy with intensity relative to the temperature of the object. The IR temperature transducer used to measure body temperature operates in the electromagnetic spectrum lying in the wavelength interval of 8 to 12 m in the far-infrared region [21]. This infrared radiation obeys all the laws of the light. These include shadowing, reflection, refraction, and other optical behavior. The infrared sensor is placed at a distance of 5 cm in front of the forehead to read the body temperature. The infrared temperature sensor we used has a fixed emissivity of 0.95. The temperature probe measures the infrared radiation and gives a dc output in millivolts. The probe has a measuring range to 260 C, with an accuracy of 3% of reading and an of output of 1 mV dc per C. The distance to spot size ratio (field of view) refers to the diameter of the spot that the probe is sensing at a given distance. The closer the object (target), the smaller the area (or spot) the probe is sensing. The relationship between spot size and the probe distance from the target is given by Spot size mm

Probe distance from the target

Fig. 3.

339

Increase in spot size with distance.

Fig. 4. Two-point calibration of weighing scale.

(1)

Hot spots can be missed if too large an area is included in the field of view, so the object should be as close as possible to the sensor (Fig. 3). The infrared temperature probe is calibrated to give a dc output in millivolts corresponding to the temperature in degrees to 260 C, Celsius. The probe has a measuring range of with an accuracy of 3% of reading and an output of 1 mV dc per C. The infrared probe is placed in one of the ears to read the tympanic temperature. The voltage output from the infrared probe is captured by the data acquisition card in the laptop, and the software stores the readings as degrees Celsius. D. Weighing Platform A Kern DE150K50 stainless-steel weighing platform balance is used to measure the body weight. The load cells are fitted on a platform and the body weight is averaged out from these load cells and recorded. The weighing range is up to 150 kg. The internal circuitry contains the signal conditioning and preamplifier to produce a noise-free voltage output. The output voltage is linear to the weight within the 0 to 150-kg range. A two-point calibration is carried out to convert the voltage reading into weight, as shown in Fig. 4. These voltage readings are sampled and stored in the laptop computer using the data acquisition card. Using calibration constants, the software will interpret the voltage into weight of a person in the range 0 to 150 kg. III. HARDWARE AND SOFTWARE The hardware system consists of several individual sensors connected to the laptop computer through the data acquisition (DAQ) card, and one sensor is connected directly to the RS232 port of the laptop computer (Fig. 1). The laptop is interfaced with a PCMCIA 12-bit resolution DAQ. External signals from physiological sensors are fed to the DAQ input channels. For each analog input channel, the DAQ card samples the signal at a

rate of 200 samples/s. This is sufficient to capture the salient features of the various signals coming in. For the ultrasound height transducer, an event counter in the PCMCIA data acquisition card is used to calculate the length of the echo pulse. The data is logged onto the laptop computer and is also transmitted to a server through a wireless local area network (LAN). The Ethernet link through the server makes it possible to view the data by any other computer connected to it and helps medical personnel to access this data via the internet for diagnosis and monitoring purposes. The system also has a self-monitoring facility so that measurements outside set limits or thresholds generate an alarm signal. The portable system has a 12-V, 3-Ah battery pack to power the transducers. Once the battery pack is fully charged, the system can be used outdoors for up to 10 h. The main program is written using National Instrument’s LabVIEW1 software and is used to log the data onto the computer. The program captures, analyzes, displays, and transmits the data to the central server. The program checks if the data directory exists on the computer and whether files may be created in the directory and written to. This is necessary to guarantee that data can be written to a file since many networks and operating systems enforce strict schemes for file protection. When the program is started, it prompts for the name or identification of the subject whose physiological parameters need to be monitored. This name is used to specify a unique file for storing the acquired data. Existing records for the same subject is untouched, and new data are appended at the end of the file every time the program is restarted. A flow chart, as shown in Fig. 5, illustrates the sequence in which the program is executed. The front panel (graphical user interface) is as shown in Fig. 6(a), and the virtual instrumentation (VI) diagram is shown in Fig. 6(b). Due to space constraints, only a part of the VI diagram screen capture is shown in the paper. Separate software modules handle each type of input. By using this approach, one is able to add different types of inputs at a later stage. The pulse 1National

Instruments, Austin, TX

340

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

by the formula as shown in (2), where the height is in meters and weight is in kilograms. BMI

Fig. 5. Flow chart to check for an existing file and to open or create a new file.

oximeter data is received by the serial-port software module, which continuously receives the heart rate and the SpO data in ASCII text format. If the sensor is not connected properly, a data error is received. The analog voltage input software module receives the weight and temperature data. The acquisition of all these signals are synchronized by the weight input, i.e., if the weight is more than 10 kg, a delay of 10 s is introduced, and then signals from all software modules are captured over a short period of 5 s. The average of this 5 s duration is taken as the final reading. This is to allow the system variables to settle before taking the readings. The counter software controller module generates a periodic pulse at a 30-ms interval to initiate the sonic bursts. When the echo pulse from the ultrasound unit starts, the counter starts counting and stops when the echo pulse ends. The number of counts will indicate the echo pulse width, which in turn corresponds to the distance between the transducer and the object. The measured distances between the transducer to weighing scale and transducer to the 2 m-pole are used for the 2-point calibration, as shown in Fig. 7. Using this calibration, the software translates the counts into distance in meters. The counter frequency is set at 100 kHz so that it can count at a resolution of 10 s. The range of the ultrasonic transducer is 3 cm to 3 m, and the corresponding echo pulse width is 100 s to 18 ms, which translates to 0.166 mm s. Hence, the software can measure the height to a resolution of 1.66 mm. The body mass index (BMI) software module calculates the BMI from the height and weight readings. The BMI is calculated

weight height

kg m

(2)

The calculated BMI is compared to the standard range to generate messages such as underweight, acceptable weight, and overweight. The data storage software module stores the data collected into the person’s records in the local drive and also transmits the data to the central server. The server displays the data through a web browser. Server Set-up On the server, the LabVIEW web server option is enabled and configured for this application, as shown in the Fig. 8. The TCP/IP option under the VI server setup in LabVIEW is disabled to avoid the configuration clash with the web server setup. The root directory for the server is configured as a shared directory. The laptop (data collection station) that is used to collect the physiological parameter data is given remote wireless access to write into this shared server directory. The shared directory is password protected to avoid unauthorized access. Inside the shared directory, a user data file is created to record the user’s name and identification number. Each time a user logs into the data collection station to record his physiological parameters, the person’s name is checked against the user data file records, and then it is decided whether the person is a new user or an existing user. For a new user, a new set of files (data file, trend graph picture file and html file to show trend graph on the web) is created. Also, for a new user, a set of lines are added to the index file in the server, these lines indicate the person’s identity and also the path for the person’s data file and trend graph file on the server. For an existing user, the previous data file belonging to the user is picked up from the server, and the new data entry is appended to the existing data file. The file is then transferred back to the server. At the same time, the LabVIEW program on the laptop generates a trend graph from the user’s present and previous data, and the picture file of this trend graph ( .png) is also stored onto the server. Hence, for a given user there are three files available on the server, a .html file to display the graph picture, a .png file that contains the trend graph picture as shown in Fig. 9, and individual data values are available in the data file .xls, as shown in Fig. 10. The trend graph is not created by the server but by the data collection station. The server only contains a picture of this trend graph, hence, zooming into individual data values is not possible on the server. The index file of the server will show up all the selectable user names whose data is available on the server. Hence, when the server is accessed by an authorized user, the person is presented with a user interface (Fig. 11), which lists the names of the persons whose records are available on the server. It also provides the option to the user to see either the trend graph or the raw data file. The Cisco systems CISCO2 AIRONET 350 series wireless LAN adapter card on the data collection station and the CISCO AIRONET 350 series wireless access point on the server side 2Cisco

Systems, Inc., San Jose, CA.

RAHMAN et al.: NETWORK APPROACH FOR PHYSIOLOGICAL PARAMETERS MEASUREMENT

341

(a)

(b) Fig. 6.

(a) Front panel. (b) Block diagram (VI diagram).

are configured for the wireless data transfer from the data collection station to the server. The maximum distance for wireless transmission between base station and server is 200 m. IV. RESULTS AND APPLICATIONS

Fig. 7. Two-point calibration of ultrasound sensor.

The data obtained from the sensors through the DAQ board is logged onto the laptop computer, and the trends are displayed, as shown in Fig. 12. It shows an actual screen capture of the subject’s physiological parameters. The graph represents height, weight, temperature, pulse rate, and SpO data, respectively. The height is displayed in meters, the weight in kilograms, the temperature in degrees Celsius, the pulse rate in number of beats

342

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

Fig. 8. LabVIEW web server configuration.

Fig. 9.

Trend graph as .png file.

RAHMAN et al.: NETWORK APPROACH FOR PHYSIOLOGICAL PARAMETERS MEASUREMENT

Fig. 10.

Spread sheet showing the data file.

Fig. 11.

Record of subjects.

per minute and the SpO in percentage, respectively. These variables were recorded for every 50 samples. This data can be sent to a server through a wireless LAN, as explained earlier. The data and trends from the server can be viewed by the medical personnel anywhere and anytime. The speed of data acquisition is much faster than the systems available on the market. Comparative studies have been made on the accuracy of the data using standard sensors available. Pulse oximeter readings were compared with a Nonin PalmSAT 2500 handheld pulse oximeter. Infrared temperature probe readings were compared with a clinical infrared ear thermometer. Height sensor readings were compared with an Evans Rule measuring tape. Weight readings were compared with a Mettler Toledo WILDCAT WS150R scale. Four subjects were selected, and for each subject the data was collected using our system and the standard devices. For each subject, three samples were taken within 10 m. The comparison data are tabulated in Table II.

343

There are some limitations. When measuring the height of a person with a tape, the height of hair is not measured, but the ultrasound transducer will measure the hair, including the hair thickness for each subject. The mean error for height is about 3 cm for subject 4, who has thick hair. The mean error for temperature is about 0.4 C, and this might have been due to positioning of the probe, as our system probe is slightly thicker than the conventional ear thermometer. The variation in heart rate and SpO are minimal compared with the standard oximeter, as both are of same brand and model with digital output. Also, these two parameters tend to vary over short periods of time. The system performance matches that of standard clinical equipment, and also, this system has the capability of transmitting data over the internet. V. POTENTIAL APPLICATIONS AND FUTURE DEVELOPMENT A. Potential Applications The complete system can be used for emergency purposes (the base unit can be fixed inside the ambulance and the sensor

344

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

Fig. 12. Screen capture of the monitoring system. The waveforms from the top represent height, weight, temperature, pulse rate, and SpO data, respectively. The amplitude is shown on the -axis and the number of samples on the -axis.

Y

X

TABLE II COMPARISON OF PHYSIOLOGICAL DATA COLLECTED USING OUR SYSTEM AGAINST THE STANDARD DEVICES

unit can be brought around), where the paramedics can attach the sensor unit to the victims on the accident site, and doctors in the hospital can monitor the victims health status and recommend the medical procedures to the paramedics online. In the case of patients, the nurse or doctor can monitor the vital signs of the patient from his office or hospital, while the patient may be moving around or may be at his home or at

some other remote location. It may also be used for teleconferencing between doctors around the globe about a certain patient’s health condition, as each doctor, irrespective of the place on the globe, will be able to see the vital parameters of the patient online. Performance-level evaluation for athletes is quite useful as the officials in charge of the athletes are not required to be in the

RAHMAN et al.: NETWORK APPROACH FOR PHYSIOLOGICAL PARAMETERS MEASUREMENT

field to monitor their performance. Instead, they can monitor it from the comfort of their office or home. As can be inferred from the categories of the users mentioned above, the physiological parameters required to be evaluated for all of them are quite similar, but their working environments are greatly varied. The emergency victims may be in an environment that requires the system to be quite rugged both in terms of communication and physical aspects. The cost of the system in such cases may not be a big factor, whereas for a tele-care patient, it certainly is. Hence, the system has to be further subdivided into smaller units so that the patients can easily afford at least the basic required units. Some of the units can be made optional for the patients who require extra functionality or extra information for themselves and can afford the extra cost. Also, a different low-cost communication system may be developed which can work in this less hostile environment and, at the same time, reduce the cost of the unit for the end user. Similarly, for the athletes, the mechanical design of the unit may be quite an important factor so that it does not interfere with their body movement and, hence, affect their performance. B. Future Development A telemetric sensor unit and a base unit for receiving data from the sensor unit for continuous monitoring are being developed in the lab. The sensor unit, capable of supporting multiple sensors, will be attached to a subject for monitoring physiological parameters, and the data from the sensor unit is sent to the base unit (receiver) through a (RF/GSM) transmitter. Three major design issues are addressed for the sensor unit. One is to develop a noninvasive, light weight, small size wireless sensor unit which can be strapped on to the body easily, the second is to minimize motion artefact, and the third is to minimize the consumption of battery power in the sensor unit. The base unit will be a stand-alone unit which will be able to receive the data from a single or multiple sensor units, store it, and also transfer it to a server unit. VI. CONCLUSION A novel portable physiological system for patient health monitoring and management with a wireless LAN has been developed. The system is suitable for good and urgent care of emergency victims, patients in hospitals or at home, and also to monitor the performance level of athletes, service personnel, and others involved in various physical activities. It is a network approach for assisting patient-centric health care management. ACKNOWLEDGMENT The authors are thankful to their Director Dr. D. Chong and Deputy Director Dr. Y. R. Huan. REFERENCES [1] R. Anderson and G. M. Lyons, “Data logging technology in ambulatory medical instrumentation,” Physiol. Meas., vol. 22, pp. 1–13, 2001. [2] B. P. Reilly et al., “A device for 24 hour ambulatory monitoring of abdominal girth using inductive plethysmography,” Physiol. Meas., vol. 23, pp. 661–670, 2002. [3] N. J. Holter, “New method for heart studies,” Science, vol. 134, pp. 1214–1220, 1961.

345

[4] E. Waterhouse, “New horizons in ambulatory electroencephalography,” IEEE Eng. Med. Biol. Mag., vol. 22, pp. 74–80, May-Jun. 2003. [5] M. Breedijk et al., “Microcomputer based system for 24-hour recording of oesophageal motility and pH profile with automated analysis,” Med. Bio. Eng. Comput., vol. 27, pp. 41–46, 1989. [6] H. Zhu et al., “A microprocessor-based data acquisition system for measuring plantar pressures from ambulatory subjects,” IEEE Trans. Biomed. Eng., vol. 38, pp. 710–714, Jul. 1991. [7] O. Toshio et al., “Human perspiration measurement,” Physiol. Meas., vol. 19, pp. 449–461, 1998. [8] K. Kario et al., “Ambulatory blood pressure monitoring for cardiovascular medicine,” IEEE Eng. Med. Biol. Mag., vol. 22, no. 3, pp. 81–88, May-Jun. 2003. [9] Petley et al., “Development and application of a general purpose ambulatory monitor,” Med. Eng. Phys., vol. 20, pp. 33–39, 1998. [10] M. G. Geoffrey, I. Ohzawa, and R. D. Freeman, “A flexible PC-based physiological monitor for animal experiments,” J. Neurosci. Method, 1995. [11] H. H. Asada et al., “Mobile monitoring with wearable photoplethysmographic biosensors,” IEEE Eng. Med. Biol. Mag., vol. 22, pp. 28–40, May-Jun. 2003. [12] S. Jayaraman and S. Park, “Enhancing the quality of life through wearable technology,” IEEE Eng. Med. Biol. Mag., vol. 22, pp. 41–48, MayJun. 2003. [13] E. Jovanov et al., “Stress monitoring using a distributed wireless intelligent sensor system,” IEEE Eng. Med. Biol. Mag., vol. 22, pp. 49–55, May-Jun. 2003. [14] J. M. Winters et al., “Wearable sensors and telerehabilitation,” IEEE Eng. Med. Biol. Mag., vol. 22, pp. 56–65, May-Jun. 2003. [15] D. Rowan, “Clinical physics and physiological measurement bibliography. Diagnostic investigations of the lower urinary tract,” Clin. Phys. Physiol. Meas., vol. 8, no. 4, pp. 379–392, 1987. [16] E. B. Phelps and T. M. Goldman, “Application of automated human voice delivery to warning devices in an intensive care unit: A laboratory study,” Int. J. Clin. Monitoring Comput., vol. 28, pp. 111–116, 1992. [17] J. C. Joseph and J. M. Brown, Introduction to Biomedical Equipment Technology. Upper Saddle River, NJ: Prentice-Hall, 2001. [18] D. P. Agrawal and Q. Zeng, Introduction to Wireless and Mobile Systems. Pacific Grove, CA: Brooks/Cole, 2003. [19] M. J. Hayes and P. R. Smith, “Quantitative evaluation of photoplethysmographic artefact reduction for pulse oximetry in biomedical sensors, fibers and optical delivery systems,” Proc. SPIE, vol. 3570, pp. 138–147, 1999. [20] H. Kocoglu et al., “Infrared tympanic thermometer can accurately measure the body temperature in children in an emergency room setting,” Int. J. Pediatric Otorhinolaryngology, vol. 65, pp. 39–43, 2002. [21] J. M. Senior, Optical Fiber Communications Principles and Practice. New Delhi , India: Prentice-Hall of India, 2001.

Fazlur Rahman (M’93) received the B.E. (Electrical) degree from Bangalore University, Bangalore, India, in 1985, the M.Tech. (Instrumentation) degree from Indian Institute of Science, Bangalore, India, in 1989, and the Ph.D. degree in electrical and electronic engineering from Nanyang Technological University, Singapore, in 2000. From 1985 to 1987, he worked as a Graduate Trainee Engineer in a private PCB manufacturing plant, Bangalore. From 1989 to 2002, he worked as a Lecturer in Bangalore Institute of Technology, Bangalore University. He joined Singapore Polytechnic, School of Electrical and Electronic Engineering in 1995, where he is currently a Lecturer, Technology leader for the Instrumentation and Photonics Group, and Manager of the Advanced Diploma Courses. He has published several papers in international journals and conference proceedings. His research interests include neural networks, nonlinear control, photonics, biomedical, and intelligent instrumentation. Dr. Rahman was awarded a scholarship to pursue the Ph.D. degree at Nanyang Technological University. He is a life member of the Instrument Society of India. He is currently serving as Secretary of Industrial Electronics Chapter, IEEE Singapore Section.

346

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 54, NO. 1, FEBRUARY 2005

Arun Kumar received the M.S. degree in computer control and automation from Nanyang Technological University, Singapore, in 1999 and the B.Tech. degree in electronics engineering from IT-BHU, Varanasi, India, in 1991. He is now a Lecturer in the School of Electrical and Electronic Engineering at Singapore Polytechnic. His main research interests lie in the area of automation and biomedical signal analysis.

Gangadharan Nagendra (S’94–M’95) was born in Jaffna, Sri Lanka, on December 14, 1963. He received the B.E. degree in electronics and communication engineering from the University of Madras, Madras, India, in 1989 and the M.E. degree in electrical and electronics engineering from the University of Auckland, Auckland, New Zealand, in 1995. From 1991 to 1994, he worked in the School of Medicine, University of Auckland, and then joined the School of Electrical and Electronic Engineering, Singapore Polytechnic. Presently, he is the Program Manager for the biomedical engineering option for two diploma programs and the Chairman of the department course management team for the Specialist Diploma in Biomedical Engineering program. Mr. Nagendra received two “Excellence in R&D” awards for his research work in the biomedical engineering area and an “Excellence in R&D” award for his work in computerized structural simulation software. He is an IEEE Power Electronics Society member and an IEEE Engineering in Medicine and Biology Society member.

Gourab Sen Gupta received the B.E. (Electronics) degree from the University of Indore, Indore, India, in 1982 and the M.E.E. degree from the Philips International Institute, Eindhoven, The Netherlands, in 1984. He is a Visiting Senior Lecturer at the Institute of Information Sciences and Technology, Massey University, Wellington, New Zealand, where he is currentlly pursuing the Ph.D. degree in intelligent control of multiagent collaborative systems. After working for five years as a Software Engineer in Philips India in the Consumer Electronics division, he joined Singapore Polytechnic in 1989, where he is currently a Senior Lecturer in the School of Electrical and Electronic Engineering. His current research interests are in the area of embedded systems, real-time vision processing, behavior programming for multiagent collaboration, and automated testing and measurement systems.