Dr. Md. Shahjahan1, Minhajur Rahman2, Abu Zahid3
2Minhajur Rahman(Khulna University of Engineering and Technology, Bangladesh)
3Abu Zahid (Khulna University of Engineering and Technology, Bangladesh)
Abstract—Existing healthcare center does not have link with online patient condition sensing system especially in Bangladesh. This paper presents design and implementation of a cheap system based on Wireless Network and HTTP protocol for the real timemonitoring of the physical condition of patient at a distance. Body sensor sense the Temperature, Pulse, Blood Pressure, Blood Glucose etc. and then the signals are processed by a microcontroller and patient data are sent by General Packet Radio Service (GPRS) shield to the web server. The healthcare center can monitor patient online and wirelessly.
Keywords—Wireless Sensor Network (WSN), GPRS, Pulse Sensor, Web based WSN.
The demand of real time as well as all day long monitoring of health status is increasing day by day. Specially it finds application in military and allied health care, monitoring patients with critical diseases like cardiac problem or diabetic, or the health condition of old people both in indoor and outdoor. Early identification of patient’s condition like critical or non-critical would help to decide whether the patient needs immediate medical attention or not. Some vital biological data like body temperature, blood glucose, pulse rate, etc. are the primary diagnostic to determine any physical disturbance or symptoms of the disease. If they are collected from the people of interest and continuously sent to a central monitoring station or server, it is possible to provide emergency service to a large number of people utilizing very limited resources. Many systems already exist to serve the purpose like popping up frequently for health care or monitoring. However, in order to remote patient monitoring there has so many system are exist regarding so many network. Wireless Body Area Network (WBAN) is a much popular system for patient monitoring. In this system some sensor nodes are connected either inside or outside of the human body, and they send their reading to a network through wireless medium. Here, at first patient’s physical data from all node sensors are gathered into a coordinator’s point and analyzed by the coordinator. Then they are sent to the health care center for further medical analysis. Usually Zigbee (802.15.4) 1 and Bluetooth (802.15.1) 2 are used to create WBAN. Although they provide a longer coverage, this kind of network required a complex arrangement of many nodes which is not economical. Heterogeneous wireless access (HWA) based remote patient monitoring system also gives some advantage over WBAN 3, as it has some unique feature like multiple wireless technologies, and this kind of system is much handier with existing communication protocols. Therefore HWA does not need any separate protocols for health monitoring. However, Wireless Sensor Network (WSN) based remote patient monitoring is much efficient than the other systems in many ways. As WSN allows wide spread mobility of every node, the doctor can monitor a patient from any place if outside of the care center 4. Even, upon integration with real time web server, the service could be made global. After reviewing some available system, we report a Web Service based remote patient monitoring system via 3G/GPRS which will be flexible and easily available system 5.
We use a compact device that acts as sensor and collect all physical data. Thereby necessity of WBAN is eliminated. Our compact device consist of a light source and detector which takes photoplethysmography (PPG) and it will help to measure some physical data like heart rate, blood pressure, blood glucose, ECG etc. 6. Using the temperature sensor the human body temperature is also measured. These physical data are initially be analyzed by microcontroller and then sent by a GPRS shield to a web server. And the data will be saved under the patient name in MySQL database for further analysis or future use.
II. Proposed System Architecture
Figure-1 shows the block diagram of proposed remote patient monitoring system. It has a compact device with microcontroller and a GPRS shield at sensor node and the GPRS shield send the data to a web server via HTTP protocol.
A compact device with temperature and pulse sensor
Microcontroller (ATmega328P) receive analog sensor data to measure physical data
GPRS shield receives data from microcontroller and sends data to web server
Web receive data from sensor node and save data into a MySQL table
Send data to Web
Figure-1: Proposed remote patient monitoring system
In sensor node, temperature sensor and pulse sensor circuits are connected with microcontroller. The microcontroller takes the analog data from the sensors, convert it into digital equivalent using it’s in built A/D converter, and finally process the sensor signal and send the measured physical data to the web server via GPRS as shows in figure: 2.
Pulse sensor circuit
ATmega32p collect data from sensor circuit and send processed data to GPRS shield
GPRS shield to send data to Web server via HTTP protocol
Figure-2: Sensing system
In the sensor node at the compact circuit, DS18B20 were used as a temperature sensor. And in pulse sensor circuit TCRT5000 were used to get photoplethysmography (PPG) for pulse measurement. The pulse circuits are established on the principle of photoplethysmography (PPG) which were used in a non-inverse method for measuring blood volume tissues using an IR transmitter as a light source and an IR receiver as a detector 7. Since the change in blood volume is related to some physical data (i.e. heart beats, blood glucose, blood pressure, etc.) via fingertip or the ear lobe 8. There are two types of PPG one is transmittance and another one is reflectance. In this paper reflective optical sensor were used which get reflectance type PPG. The reflected type PPG the light source and detector were placed in the same side when the light source emitted light some light transmitted and remain part were being reflected.
The PPG signal has two components, AC and DC. The DC part of the average blood volume in the tissues and AC part of the variation of blood volume in the tissues. So, by removing a DC component from the PPG signal it is possible by analyzing AC component to measure heart rate as well as blood pressure and blood glucose.
B. URL and MySQL data table
GPRS shield sends sensor data to a predefined URL. The URL consists a PHP page and programmed with HTTP GET method. So, this page receives data and save this data to a predefined table at MySQL under the patient name is shown in the figure-3. And the saved data are prepared to monitor or further analyze by medical personnel.
Web page of the predefined URL
Save to MySQL
MySQL table to save data by patient name and data name
Figure-3: Web Server
III. Experimantal Setup for WSN
The experimental setup has been constructed in three steps: sensor circuit, GPRS shield interfacing with microcontroller and preparing Web server to data retrievals.
A. Sensor Circuit
A compact device was designed to measure physical data, in this circuit there a temperature and pulse were measured. Temperature sensors DS18B20 were used and the outputs of temperature sensor collected by the Arduino Uno board and measure temperature in °F scale.
For pulse sensor circuit, TCRT5000 were used as a reflective optical sensor, which has an IR transmitter as light emitter and phototransistor as a detector. As a reflective optical sensor, it will give reflected type PPG. Sensor given output contain AC and DC component and there has two stage filtering arrangement to remove the DC component and amplify the remaining AC component.
In figure-4, an external biasing circuit was used to enable the TCR5000. The transistor BC547 were connecting with the sensor part when the base of the transistor is getting higher the sensor will enable. The phototransistor receive reflected light by tissues and as output it gives variable voltage which has AC and DC component, in the figure-5 the output of detector goes to the first stage filter. The first stage filtering starts with a passive low pass filter to remove DC component. And the cutoff frequency of passive LPF has been settled at 0.7Hz by choosing 4.7µF capacitor (C) and 47K? resistor (R) shown in figure-5. The Active high pass filter amplifies the remaining AC component using op-amp LM358 9. The gains were settled at 101 to amplify the remaining AC component by choosing resistor R1=680K? and R2=6.8K? shown in figure-5. Also the cut-off frequency of active HPF is 2.34Hz.
The 101 times amplifying signal were imported to second stage filtering and after passive LPF and active HPF the output after second stage will be 101×101=10201 times of the original signal. The two stages of filtering and amplification converts the input PPG signal to near TTL pulse and they are synchronous with the heartbeat. The frequency of these pulses is related to the heart rate (BPM) 10. And these TTL pulse signals collected by ATmega328P and calculate the BPM using the following algorithm 11 in figure-6.
Figure-4: External Biased Circuit
Figure-5: Filter Circuit
B. Data Sending via GPRS
The special feature of GPRS technology is the worldwide coverage. So, the health care will not be limited by any kind of distance. And any kind of physical data or signal (i.e. ECG) can be transferred in packet format via GPRS 12. In the circuit, the SIM900 Seeedstudio GPRS shield was connected with ATmega328P via Arduino UNO board and the finalized data send via GPRS shield to the web server.
The GPRS shield can be configured with AT command after interfacing the shield with PC via UART port 13. Figure-7 shows the sequence of configuring GPRS shield by AT command.
Read 600 consecutive ADC samples at 5 ms interval
Apply a 10-pont moving average filter to smoothen the PPG signal
Computer heart rate based on three successive peaks in the PPG waveform
Display heart beat rate
Remove DC component
Scale ADC samples to 1-1023
Check if ADC samples range is enough (?50). If not the input data is invalid (noise) and repeat above steps
Figure-6: BPM calculating Algorithm
Check the GPRS service state(if Attach or Detach)
Initialize HTTP service
Send data to predefined URL
Describe HTTP method (i.e. GET)
Set the type of internet connection (i.e. GPRS)
Send request for IP address
APN matching (Access Point Name)
Figure-7: Steps of GPRS shield configuration.
C. Database creation in MySQL
A web server has been properly arranged to get data by HTTP GET method 14. A predefine URL (i.e. abdc.php?) was used to send data at GPRS side and the same variable were declared in the given URL page. Then the PHP code is written on the basis of GET method and variable to save the received data in MySQL server table 15.
IV. experiment and result
In the figure-9, the experimental setup of compact device has been shown. The experiments were accomplished by three steps: Temperature measurement and comparison with actual value, Pulse measurement and comparison with actual value, and Send data to the web server via GPRS shield.
A. Temperature Sensor
In the TABLE-2 experimental temperature of five people and actual temperature measured by clinical thermometer being compared.
TABLE-2: Temperature Sensor
Percent of Accuracy
From the Percentage of Accuracy column it is inevitable that the temperature sensor work perfectly and this sensor is preferable for WSN.
B. Pulse Sensor
In the TABLE-3 experimental pulse of five people and actual pulse measured by wrist being compared. The actual pulse was measured from wrist by counting 15seconds pulse, then multiply the result by 4 and this process gives pulse in Beat Per minute (BPM).
TABLE-3: Pulse Sensor
Pulse sensor circuit
Percent of Accuracy
From the Percentage of Accuracy column, it’s clear that the pulse sensor circuit and algorithm is quite perfect. So, this kind of setup is preferable for WSN.
C. Remote monitoring
The compact devices were placed on the hand and the TCRT5000 and DS18B20 were mounted on the finger. After powering the circuit microcontroller start measuring physical data with the help of the physical sensor then the measured data can be seen via serial monitor via the Arduino UNO board. After few second of powering the circuit, the web server starts responding. In the figure-8, some experimental data on MySQL table on the web server has been shown.
Figure-8: Data monitoring (at MySQL)
From the data table, the medical personnel can monitor remote patient from any kind of distance.
Figure-9: Experimental Setup of compact device.
From the above approach it’s clear that, this proposed system reduce the need of many sensor nodes by sensing more than one physical data from one body part which save the cost of sensor node circuit. Also many sensor nodes require more wireless device.
V. Cost analysis and Comparison with existing system
Roughly this system costs 5000BDT 16, 17 at experiment, but other existing systems presently at market cost more since more sensor nodes have been applied in those systems. As a result, this wireless device’s cost has been optimized in this system. Moreover, this system cover a large area as a matter of fact it can be established from one country to another so, patient’s physical condition can be monitored by doctor at hospital even if he/she is residing at home within the same country or another .This feature is beneficial when the patient is discharged recently from hospital but still need some attention from doctors.
We developed a compact device to monitor the health condition of patients from remote place. The compact device produces results which are in an acceptable range. The proposed system worked successfully, compared to the other available system this is a cost-effective solution to remote patient monitoring. However, with a few modifications the system can be made a bit more reliable and more efficient. Currently in pulse sensor circuit path optical sensor must need to keep very still with the body part and the reflective optical sensor must need to protect form surrounding visible light. Studies have shown that reflective optical sensor can also use to measure blood pressure, blood glucose and respiratory rate. If further studies provide an acceptable result for above physical data using reflective optical sensor. It might be accomplished using the same basic circuit with some modification. With some fine tune, the compact device can be made as a perfect sensor device to measure possible more physical data in a cost-effective way. And it will help to avoid more sensor node cost by one sensing node for all physical data sense.
1 Sharma, R. ; Gupta, S.K. ; Suhas, K.K. ; Kashyap, G.S. “Performance Analysis of Zigbee Based Wireless Sensor Network for Remote Patient Monitoring,” in Fourth International Conference on Communication Systems and Network Technologies (CSNT), April 2014, pp. 58 – 62.
2 Vyas, A.L. ; Thakker, B. ; Mulvaney, D. ; Datta, S. “Wireless Body Area Network for health monitoring,”on 4th International Conference on Biomedical Engineering and Informatics (BMEI),Oct.2011, vol.4, pp. 2183-2186.
3 Niyato, D. ; Hossain, E. ; Camorlinga, S.”Remote patient monitoring service using heterogeneous wireless access networks: architecture and optimization,” IEEE Journal on Selected Areas in Communications, vol. 27, issue. 4, pp. 412-423.
4 Poenaru, E. ; Poenaru, C.”Networking Architectures for Healthcare Wireless Sensor Networks,” on 11thRoedunet International Conference (RoEduNet), jan, 2013 , pp. 1-6.
5 Eddabbah, M. ; Moussaoui, M. ; Laaziz, Y. “A flexible 3G WebService based gateway for wireless sensor networks in support of remote patient monitoring systems,” 14th Mediterranean Microwave Symposium (MMS), Dec. 2014, pp. 1-5.
6 NazmusSaquib, Md. Tarikul Islam Papon, Ishtiyaque Ahmad, and AshikurRahman, “Measurement of Heart Rate Using Photoplethysmography,” International Conference on Networking Systems and Security (NSysS), Jan. 2015, pp. 1-6.
7 Chi Kin Lao, U Kin Che, Wei Chen, Sio Hang Pun, Peng Un Mak, Feng Wan, and Mang I Vai, “Portable Heart Rate Detector Based on Photoplethysmography with Android Programmable Devices for Ubiquitous Health Monitoring System,” 35th International Conference on International Science and Engineering Society, July 2012, pp. 605-609.
8 Monte-Moreno E, “Non-invasive estimate of blood glucose and blood pressure from a photoplethysmograph by means of machine learning techniques”, Artificial Intelligence in Medicine 53 (2011) 127– 138
9 LM358 (ACTIVE). Online. Available: http://www.ti.com/product/lm358.
10 Claudio Becchetti, Alessandro Neri, Medical Instrument Design and Development, equation 2.39
11 EASY PULSE METER. Online. Available: http://embedded-lab.com/blog/?p=8183
12 GPRS Shield V2.0. Online. Available: http://www.seeedstudio.com/wiki/GPRS_Shield_V2.0
13 SIM900 AT Command Manual. Online. Available: http://www.seeedstudio.com/wiki/File:SIM900_AT_Command_Manual_V1.03.pdf
14 HTTP Methods: GET vs. POST. Online. Available: http://www.w3schools.com/tags/ref_httpmethods.asp
15 The GSM/GPRS & GPS Shield: some Http connections examples. Online. Available: http://www.open-electronics.org/the-gsmgprs-gps-shield-some-http-connections-examples/
16 SIM900 GSM/GPRS Shield Online. Available: http://projuktiponno.com/SIM900-GSM-GPRS-Shield-projuktiponn-bangladesh
17 Arduino Uno R3 (China) Online. Available: https://www.techshopbd.com/product-categories/boards/1253/arduino-uno-r3-china-techshop-bangladesh