Web Server Based Remote Health Monitoring System István Bosznai, Ferenc Ender, and Hunor Sántha Department of Electronics Technology, Budapest University of Technology and Economics, Budapest, Hungary, 1111 Budapest, Goldmann Gy. t. 3., building V2 Phone: +36 1 463 3730, Fax: +36 1 463 4118, [email protected]

Abstract: In this paper a new solution of a home monitoring system is presented. Home monitoring makes possible for the patient to measure different physical parameters at home, and for the physician to check the results anywhere without personal meeting. Contrary to conventional home monitoring systems, the realized system uses distributed data storage of patients’ data, instead of using a remote server to store all the data. A fully functional home monitoring system has been realized, that contains a microcontroller based web server to store patient data. This unit collects data via Bluetooth from a small size wearable Electrocardiograph (ECG) device designed and constructed by the authors. The size of the realized web server is 9·11·3 cm - w·l·h, and the power consumption is only 2W. The stored data can be accessed via internet. The remote client runs a Java application stored on the microcontroller based web server. The physician uses this Java application to access and view patients’ data in a remote location and to form a diagnosis.

1. INTRODUCTION Home Health Care means any type of care given to a person in their own home. Home Health Care aims to make it possible for people to remain at home rather than use residential, long-term, or institutional-based nursing care. Home Monitoring is a subsection of Home Health Care that uses a remote device to monitor the vital signs of a patient. This vital sign can be for example the temperature, oxygen saturation, heart rate, etc. of the patient. A typical home monitoring system consists of two parts: a device at the patient’s location that records the patient data, and a device that displays the patient data at the physician’s office. Diseases of the heart and cardiovascular system (cardiovascular disease or CVD) are the main cause of death in EU: accounting for over 2 million deaths each year. Nearly half (42%) of all deaths in the EU (46% deaths in women and 39% deaths in men) are from CVD - slightly less than for Europe as a whole. Between a third and a half of deaths from CVD are from Coronary Heart Disease (CHD) and around a quarter are from stroke. CHD by itself is also the

single most common cause of death in the EU: accounting for over 741,000 deaths in the EU each year. Around one in six men (16%) and over one in seven women (15%) die from the disease. [1] Electrocardiogram is the recording of the electrical activity of the heart over time via skin electrodes. It is a noninvasive recording produced by an electrocardiograph. The impulses from the sinoatrial node stimulate the muscle fibers of the heart to contract. The generated electrical waves can be measured at selectively placed electrodes (electrical contacts) on the skin. Electrodes on different sides of the heart measure the activity of different parts of the heart muscle. An ECG displays the voltage between pairs of these electrodes. [2] Using an ECG recorder in Home Monitoring System cardiac arrhythmias, including fibrillation, bradycardia, and tachycardia can be detected. This system also can be used to monitor post heart-attack status of a patient, early recognition of unexpected lesions occurring invasive cardiological intervention is possible by such a device as well. Another area of this Home Monitoring system is the early recognition of dysfunction of the pacemaker in pacemaker

wearing patients and the continuous monitoring of newly implanted pacemaker patients. [3] A Home monitoring system that records ECG at the patients’ home can be used to monitor almost all CHD, arrhythmias, and post heart-attack events, which means a huge market potential. Our Research and Technological Development (RTD) work aimed to create a small size, easy to use, low cost, low power consumption unit that can replace the centralized high power consumption database servers, and can mean an other approach, instead of the actually proposed, centralized e-Health systems. By reducing the power consumption and making the patient data distributed the cost of establishing a home monitoring system could be dramatically reduced.

2. EXPERIMENTAL 2.1. Literature survey Before creating the Home Monitoring System, a literature survey has been done using ScienceDirect, database of European Patent Office and Google to find out whether such device exists in the market or not. During the search 13 systems were examined from different aspects (size, price, GSM connectivity, internet connectivity, multiple recording units, connects to remote server, uses distributed data storage) with the following results: 9 of the 13 systems connects to the physician through the internet, 6 of the 13 provides accessibility and display of the measurements via World Wide Web, and only 3 systems support multiple devices, not only an ECG, and all the examined systems connect to a remote server to store the patient data, none supported distributed data storage. The survey resulted that developing a Home Monitoring System that uses distributed data storage, handles multiple devices, is small, and power saving may be a useful alternative in certain applications. 2.2 Hardware development methods The schematic diagram and the layout of the Web Server was created in OrCAD 15.7 and was designed to fit in a standard size box. The size of the whole Web Server is quite small (9·11·3 cm - w·l·h). The main part of the Web Server is a PIC18F67J60 microcontroller from Microchip Inc. that contains an

embedded Ethernet control module. The storage is a Kingston 2 GB memory card; the EEPROM is a 256 Kbit I 2C module from Microchip. The display unit is an EA-DOGM 162E Liquid Crystal Display (LCD) that can be accessed via SPI. The Real Time Clock (RTC) is made by Maxim, and it also communicates via the SPI bus. The Bluetooth module is a WT12 module that is produced by Bluegiga. 2.3 Software development methods The firmware of the Web Server unit was written in C programming language, and the compiler that was used is the CCS C from Custom Computer Services, the compiler is used in Microchip’s MPLAB GUI. The firmware contains pre-written OEM libraries such as the SD card handling routines, or the TCP/IP stack that controls every Ethernet communication, including the Web Server module, every other library had to be written, these include the library for the EEPROM, LCD, and the RTC. Every self created library consists of three parts: initialization, data reading part and data writing part. The application on the remote client was written in Sun’s Java that derives much of its syntax from C and C++ but has a simpler object model and fewer lowlevel facilities. Another advantage of the Java application is that besides the universal Java Virtual machine (that required viewing lots of web pages) nothing needs to be installed on the client computer. 2.4 Holter ECG as sensor input The ECG recorder was developed by the authors in 2006 as a Student’s Scholarly Circle. [4] The device records the signal of two standard Einthoven leads, EI and EII. The 1 mV ECG waves are amplified 1000 times so the microcontroller samples an 1 V signal (Fig. 1.). The sampling frequency of the signals is 500 Hz and the resolution is 10 bits, which means 1024 steps, using 3.3 V supply voltage the smallest voltage increase/decrease that the recorder can detect is 3.2 mV. The device communicates via Bluetooth, and the communication speed is 115200 bits/sec (11.2 Kbyte/s). The data transmission speed is 1.95 Kbytes/sec, this amount data is easily transmitted with the 11.2 Kbyte/s maximum bandwidth.

3. RESULTS 3.1. The system plan

Figure 1: The schematic diagram of the ECG recorder

The ECG recorder uses a Lithium-ion battery as the main power source. The capacity of the battery is 1100 mAh, the total current consumption of the device is 32.3 mA, so the ECG can operate 34 hours continuously.

The aim was to create a system that can record biological signs of a patient, can be accessed remotely, is application specific, small in size and low power consumption, easy to use, in terms, that the system automatically sets the settings. After creating the diagram (Fig. 3.) of the system the components were chosen.

The complete ECG recorder can be seen in Fig. 2. The coloring of the electrode connectors is according to the European ECG color standard.

Figure 3: The system plan for the Home Monitoring System

Figure 2: The ECG recorder with the disposable 3M electrodes

The red electrode goes to the right arm, the yellow electrode goes to the left arm, the green electrode goes to the left leg and the black electrode goes to the right leg. 2.5 Validation methods The system was tested on almost 40 healthy volunteers, including young and adult people from both sexes. The web server recorded, and stored their data successfully; the measured data were accessible via the Java application. Five measurements were taken on a day, after all the measurements were done the Web Server was left at the university and a cardiologist evaluated all the measurements from a remote location.

For the Human sensor an ECG device was chosen because of the reasons mentioned in the introduction part. The device that controls the ECG is a Web Server, this device ensures two things: the data storage, and the remote accessibility of the data. The communication between the ECG recorder and the Web Server is established by Bluetooth. Because small size and low cost is required a microcontroller was chosen to realize the web server. The remote interface at the physician’s location is a Personal Computer (PC) that runs a Java applet stored on the microcontroller based Web Server module. 3.2. The realized Home Monitoring System The system consists of two hardware and one software parts (Fig. 4.): 

An ECG recorder which records physiological signs of the patient



A web server which stores the data, and communicates with the ECG unit via Bluetooth, and the Remote Client via TCP/IP

the



A Java application which displays the recorded ECG on the remote PC

master/slave mode where the master device initiates the data frame. There are four SPI enabled parts in the Web Server unit. The first is the EEPROM that stores data even after the device is turned off, these data includes the current date, setup settings, MAC address of the unit, etc. The next SPI unit is the LCD that displays information to the patient, such as the date, the IP address of the unit, or the status of the measurement.

Figure 4: The schematic diagram of the home monitoring system

While the patient makes the measurement the Web Server unit stores the data locally on an SD (Secure Digital) memory card. The realized Web Server also communicates with the recording unit, initiates-, stops the measurement, and displays the data on a LCD. The communication protocol is Bluetooth, so not only an ECG unit can be connected to the Web Server unit, but e.g.: a Photoplethysmograph (PPG) or Blood Pressure (BP) meter having a Bluetooth data link. The physician accesses these data remotely, using internet. The Java application, which displays the ECG curves, patient data and several other properties, is stored locally on the web server, so the physician needs only a web browser to view the results. 3.2. The web server and data storage unit The Web Server can be separated into seven parts (Fig. 5.). The chosen microcontroller has an embedded Ethernet controller, which is a complete connectivity solution, including full implementations of both Media Access Control (MAC) and Physical Layer transceiver (PHY) modules. Two pulse transformers and a few passive components are all that are required to connect the microcontroller directly to an Ethernet network. This family introduces a new line of low-voltage devices with the foremost traditional advantage of all PIC18 microcontrollers namely, high computational performance and a rich feature set at an extremely competitive price point. These features make the selected microcontroller family is a logical choice for many high-performance applications where cost is a primary consideration. [5] The microcontroller manages the connection to other Bluetooth devices; it controls the devices via the Serial Peripheral Interface Bus (SPI bus). It is a synchronous serial data link standard that operates in full duplex mode. Devices communicate in

Figure 5: The schematic diagram of the Web Server

The third unit is the SD memory card that holds the patient data such as birth date, identification, etc. and the measurements. A Kingston 2 GB SD card can store approximately 12 days of data. The SD card was chosen of the many memory card types because it can be accessed via SPI witch almost all the peripherals use. The maximum data transfer that can be achieved with the SPI mode is 25 Mbit/s. A 24 hour measurement with the 1.95 Kbyte/sec transmission would take 164 Mb, this could be transmitted in 52 seconds to the client PC. The last component is a Real Time Clock that keeps track of the current time. This is needed to add timestamp to every measurement. Most RTCs use a 32.768 kHz crystal oscillator, this frequency is exactly 215 cycles per second, which is a convenient rate to use with simple binary counter circuits. The Bluetooth transceiver and the USB communicate via serial connection that is implemented in the microcontroller.

A low power Class II (10 meters range) low-cost transceiver chip was selected. Bluetooth makes it possible for these devices to communicate with each other when they are in range. [6]

consumption of a PC based web server exceeds more than 300 W.

3.3. The firmware of the Web Server unit The main program first initializes the peripherals, like the SD card, EEPROM, the LCD, reads the current date, initializes the Ethernet module, etc. after that Web Server is ready string is written to the LCD. When the initialization is done the program enters the Web Server state, where it listens to incoming http requests. This is the state where the physician can remotely access all measurement data. Meanwhile the Bluetooth module is continuously searches for the ECG recorder, when it is connected the main program changes to measure mode and the Web Server starts the measurement. The unit does not write the received raw data to the memory card directly, but uses a FIFO to store the incoming data while the memory card allocates new space, and cannot accept data. After the measurement is done the Web Server turns the ECG recorder off, closes the measurement file, sets the new parameters and the program gets back to the web server mode. (Fig. 6.)

Figure 6: The block diagram of the firmware

3.4. The realized Web Server The completed Web Server unit can be seen in Fig. 7. The power consumption was measured, the unit consumes 0.55 A, the supply voltage is 3.3 V, this gives a total power consumption of 0.55 A·3.3 V = 1.815 W. By reducing the power consumption and making the patient data distributed the cost of establishing a home monitoring system could be dramatically reduced. The average power

Figure 7: The completed Web Server unit

3.5. The application on the remote client The aim was to create an application that can be run on any operating system.

Figure 8: The Java application at the remote client with the recorded ECG signals

The whole application is stored on the Web Server unit making the measurements accessible almost everywhere, and compatible with every PC. The application organizes the measurements according to date (Fig. 8.). It was mentioned previously that the ECG recorder records only two standard Einthoven leads, but the application draws six leads. This is possible because only two of the six leads are independent; the other

four can be easily calculated. The formulas for the other four leads are: EIII EII  EI

EI  EII 2 EI  EIII aVL  2 EII  EIII aVF  2

aVR

(3.2)

As we can see the patient only have to put the electrodes on, and turn on the ECG recorder, besides that the whole measurement is automatic. After the measurement is done the physician can access the data.

(3.3)

4. CONCLUSIONS

(3.1)

(3.4)

Leads aVR (augmented vector right), aVL (augmented vector left), and aVF (augmented vector foot) are augmented limb leads. They are derived from the same three electrodes as leads I, II, and III. However, they “view” the heart from different angles. The recorded ECG signals are superimposed with noise for example 50 Hz and muscle movement noise. A Finite Impulse Response (FIR) low pass filter is realized in the Java application to filter the previously mentioned noises added to the signals (Fig. 9.).

During the development a fully operational Home Monitoring system was realized, that can be used to monitor various heart diseases. The system uses a wearable ECG to monitor the heart activity of the patient. The whole measurement process is fully automatic, the only task of the patient is to put the ECG electrodes on and turn the ECG device on. The remote Java application displays the measurement data. The application is stored on the Web Server unit, thus the physician needs only a web browser to display the results. A complete test including 40 healthy volunteers has been taken out, where the system recorded and displayed their ECG successfully.

REFERENCES

Figure 9: The ECG signal without and with filtering

3.6. The measuring process The measurement is fully automatic; it can be divided into 5 steps: 

The patient places the ECG electrodes to the arms and the legs



The patient turns the ECG recorder on



The Web Server connects with the ECG recorder



The measurement takes as much time as the physician previously defined



The Web Server closes the connection to the ECG module and turns it off

[1] CVD mortality in Europe http://www.heartstats.org/datapage.asp?id=754 [2] The Complete Guide to ECGs by James O'Keefe, Stephen Hammill, Mark Freed, and Steven Pogwizd (Paperback - Oct 3, 2008) [3] Ede Kékes, “The real value of the transtelephonic ECG system in the clinical cardiological practice”, vol 148, (issue 31), pp. 1443-1451, Aug. 2007. [4] István Bosznai, Zoltán Kovács, “Design and Realization of a Wireless ECG Processing and Recording System” Student’s Scholarly Circle paper, 2006 [5] A datasheet of PIC18F67J60 microcontroller http://www.microchip.com/wwwproducts/Devices.aspx?dD ocName=en026445 [6] Newton, Harold. (2007), Newton’s telecom dictionary. New York: Flatiron Publishing.