Wireless Biomedical System Design Based on ZigBee Technology for ...

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09)

MUSCAT, FEBRUARY 15-18, 2009

Wireless Biomedical System Design Based on ZigBee Technology for Autonomous Healthcare Nabil Hamza, Farid Touati and Lazhar Khriji Department of Electrical & Computer Engineering, Sultan Qaboos University, Muscat, OMAN. [email protected], [email protected], [email protected]

Abstract—We present a framework for a wireless health monitoring system using ZigBee technology. It has the capability to monitor vital signals from multiple biosensors. Biomedical signals are collected and processed using 2-tiered subsystems. The first stage is the mobile device carried on the body that runs a number of biosensors (internal subsystem). At the second stage, further processing is performed by a local base station (external subsystem) using the raw data transmitted on-request by the mobile device. The raw data is also stored at this base station. The processed data as well as the analysis results are then monitored and diagnosed through a human-machine interface. The main advantages of the proposed framework are (1) the ability to detect signals wirelessly within a body sensor network, (2) low-power and reliable data transmission through ZigBee network nodes, (3) secure transmission of medical data over a body sensor network, (4) efficient channel allocation for medical data transmission over wireless networks, and (5) optimized analysis of data using an adaptive hardware architecture that maximizes the utility of processing and computational capacity at each platform.

systems face a number of challenging tasks, as they need to address often quite conflicting requirements for size, operating time, precision, and reliability. Lately traditional system to collect parameters for daily homecare is widely used in biomedicine. The traditional system adopts wired way wiring which makes the system complex, bulky and expensive, fig.1. Adopting wireless way wiring is convenient and economical. Wireless biomedical sensors are a group of embedded smart sensors that form a network from wireless communication links [3] and operate within the human body to compensate for various diseases. The smart sensors are placed in the body (in-vivo monitoring). They detect condition and digitize physiological signals change in its environment and communicate with a base station via a wireless interface as it is impractical to distribute wires throughout the body, due to its complexity and limitation of subject's motion [4], fig.1. Applications using or taking interest in wireless biomedical sensors are artificial retina, glucose level monitors, organ monitors, cancer detectors, and general health monitors to name a few.

Keywords: Vital signals; ZigBee; Autonomous healthcare; Adaptive architecture; Reliable and secure data-transmission.

I.

INTRODUCTION

As numerous wireless personal area networking (WPAN) technologies emerge, the interest for applications such as health monitoring, smart homes, and industrial control has grown significantly. ZigBee is the first industrial standard WPAN technology [1] that provides short-range, low-power, and secured communication, and supports mesh networking and multi-hopping. It is a new wireless network protocol stack of IEEE 802.15.4 for use in industrial equipment and home appliances in order to take in multi-type, multi-point sensor information [2]. While many smart home application areas such as lighting, security, and climate control have been suggested using the ZigBee standard, health-care applications have not received much attention despite their importance and high-value added. Here, we present a wireless communication system for real-time health monitoring with secure transmission capability. One of the most promising applications of sensor networks is for human health monitoring. A number of tiny wireless sensors, strategically placed on the human body, create a wireless body area network that can monitor various vital signs, providing real-time feedback to the user and medical personnel. The wireless body area networks promise to revolutionize health monitoring. However, designers of such © SQU-2009 ISSN: 1813-419X

Figure 1. Comparison between old Biomedical System (Wired in-vitro biosensors) and New Biomedical System (Wireless in-vivo biosensors)

With such a smart system including multiple sensors of different types, we envision a future where biosensors can form a wireless sensor network, as dot matrix sensors, comprising a large number of nodes whose placement in the body can be either pre-determined or random according to the application [5, 6]. Since achieving self-powered of implantable sensors is difficult with current technology, biosensors must operate with very limited power. Prior research shows that the pulsed mode for complete measurement system is the lowest power consumption compared to the continuous measurement mode [7]. Pulsed mode is based on the idea that energy consumption can be reduced by having only a small fraction of nodes perform long range communication with a base station, compared to the continuous mode.

183

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09)

Our aim is to develop a reliable multipurpose prototype solution, providing all the above-mentioned functionalities to prove the feasibility of the integration of the whole system in a compact system-on-chip (SOC) [8, 9, 10]. With such a generic fully programmable IC, the cost-efficient development of biomedical devices for different applications in the field of homecare and emergency comes within reach [11]. In a first phase, we describe both the implementation of the acquisition prototype with wireless transmission capability (short-range ZigBee module); and the tool of real-time acquisition, processing, storage and visualization in the base station unit.

MUSCAT, FEBRUARY 15-18, 2009

It is a short range technology that allows secure and robust communications. The use of radio device, capable to transfer data over a range of up to 100 meters outdoor-line of sight and up to 30 meters indoor-urban, is well recommended. As user interface, a windows API (Application Programming Interface) application is to be developed with high level design software for graphical user interface development. B. Development environment The conception, design and implementation of the entire acquiring prototype were carried out using PROTON® Basic Compiler Development Suite (Crownhill Associates), PROTEUS® Professional (Multipower Technology), and EAGLE® Layout Editor (CadSoft Computer). The graphical user interface was created under LABVIEW® (National Instruments).

The outline of this paper is as follows: section II describes the system analysis, the employed technology and the development environment. In section III and IV, hardware and implemented software will be explained, respectively. Experimental results and final prototype are described in section V, while section VI concludes the paper.

Figure 2. Overall block diagram of the acquiring chain

II.

SYSTEM ANALYSIS

III.

Our novel model of biomedical implant is designed with the following technical features: • Large versatility (Micro-controller embedded system). • Efficient power management (power saving technique). • Pulsed measurement mode concept. • Embedded interface with a dot matrix sensors (front-end). • On request addressable sensor node (polling mode). • Selectable data refreshment time. • In-situ wireless data programming and upgrading (bootloader technique). • Bidirectional data link (half duplex with reverse telemetry). A detailed description of the whole system will be discussed in the following sections.

The overriding challenge in developing such a biomedical implant is to keep them small in size and economical in energy consumption but with a high flexibility in terms of adapting them to various measurement types and also with ease of implantation. To overcome these constraints a basic architecture has been designed, with an embedded MCU core which is not variable (hardware architecture) in all measurement types. The block diagram of the whole system is shown in Fig.2. It consists of two main prototype parts: Testbed prototype part and Implant prototype part. A. Implant’s prototype The suggested architecture has the minimum requisites of an autonomous system: a dot matrix multi-sensors array, an intelligent power saving and supervision, an embedded microcontroller and a serial communications block for the twoway transmission of data (half-duplex) through a ZigBee RF transceiver module. The latter allows the MCU to receive instructions, memory program upgrade and gather information from requested sensors (polling mode).

A. Selected Technology We have chosen a ZigBee RF Module to meet IEEE 802.15.4 standards and the ISM 2.4 GHz frequency band. XBee RF Module complies with Part 15 of the FCC rules and regulations [12]. It supports the unique needs of low-cost, lowpower wireless sensor networks. The module requires minimal power and provides reliable delivery of data between devices.

© SQU-2009 ISSN: 1813-419X

SYSTEM OVERALL ARCHITECTURE AND DESIGN

184

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09)

Microchip’s PIC18F4550 Microcontroller is selected because it has an On-chip FLASH Program Memory with In-System Programming (ISP) which makes it flexible and easy to upgrade. It is an 8-bit RISC (Reduced Instruction Set Computer) circuit with 8 A/D converters, 3 timers, serial UART (Universal Asynchronous Receiver Transmitter) and an enhanced USB interface working with a clock frequency up to 20 MHz. Its internally implemented RISC architecture gives an instruction runtime of between 80 and 200 ns depending on the chosen oscillator.

Communication between the implant’s and the Testbed’s prototypes is achieved via a serial channel needing only the Tx and Rx lines which are wirelessly transmitted. Compared to previous designed systems [10, 11, 13, 14], the proposed architecture presents some significant advanced features such as: (1) an optimized read-out architecture for multi-channel implant (single conditioning bloc for all distributed sensors), (2) a windows API graphical user interface, (3) a wide range of programmable parameters (sensor category, sampling time, data transfer bit rate, etc.), (4) a realtime physiological signals monitoring with selectable mode (continuous, single and polling mode), etc.

The dot matrix sensors array contains several sensors, which can sense one or more physical quantities. The addressable sensor interface chip provides the address, the amplification and analog-to-digital conversion of the sensed signal. It contains an analog multiplexer, a programmable analog front-end and a ten bits analog to digital converter, makes the sensor interface chip a versatile component, which can be programmed at any time. It offers also, options for intelligent power management. Indeed, all channels which are not in use can be switched off individually. Thus, the microcontroller has several important tasks:

IV.

SYSTEM CODESIGN AND IMPLEMENTATION

The software development can be divided into two levels:

(1) It controls the sensor interface chip and provides its settings, such as the configuration of the readout electronics like sensor address and analog front-end configuration as well as sensor-specific software routines.

ƒ

Low level software microcontroller.

ƒ

High level software for the applications in the Base-Station (Graphical User Interface).

associated

to

the

A. Firmware implementation (MCU) The communication protocol between the external unit (Testbed’s subsystem prototype) and the internal unit (Implant’s subsystem prototype) can be depicted by the flow charts illustrated in Fig.3 and Fig.4. Both microcontrollers’ programs have been structured into two parts as given in Fig.5:

(2) It gathers the data coming from the sensor interface chip and stores it in a memory. (3) It implements some smart compression algorithms (base-band coder/decoder) to reduce the energy consumption during data transmission. The bi-directional communication link sends the sensor's data to the transceiver and provides the microcontroller with new programming instructions. Hence, the accuracy, the sensitivity, the acquisition rate and the data processing can be changed during operation, which are necessary to adapt the system to the environment changes and to compensate for drift phenomena.

ƒ

Main program: It configures the MCU settings and stands-by waiting for interrupt event.

ƒ

Interrupt subroutine: It is a program that runs only when an interrupt event is arrived. Hence, it is not usually activated and is brought into action only by a predetermined interrupt event.

The Testbed’s MCU is the main component of the whole acquiring chain. Indeed, it has two interrupt vectors as shown in Fig.3: ƒ

One interrupt enabled when receiving an instruction from the base station through its incorporated USB interface and,

ƒ

Another one enabled when receiving sensed data wirelessly from the implant’s MCU through its implemented USART.

B. Testbed’s prototype One of the main components of our system is the Testbed subsystem, which performs all functionalities fixed above. It consists of Graphical User Interface (GUI), USB interface, Microcontroller and ZigBee RF Module.

Once data flow is received from the implant’s MCU, the Testbed’s MCU has other tasks to execute such as: data storage, computing and transmitting to the base station before standing-by once again waiting for another acquisition request.

The GUI interface is a windows tool with a user interface that performs dialogue with the patient's implant and offers the following functionalities: category settings of health diagnosis, data refreshment time selection and signal reception; visualization and storage.

The implant’s MCU has minimum tasks to run. In fact, its tasks are limited at standing-by waiting for an interrupt coming from receiving instructions sent wirelessly by the Testbed’s MCU through its USART. Once instructions are received from the Testbed’s MCU, the implant’s MCU decodes the next incoming eight bits, sends them to the address bus (to multiplex the requested sensor) and gathers the digitized-recorded data for being sent to the outside (polling mode) and stands-by again waiting for another request.

The heart of the Testbed is a microcontroller unit MCU, which provides stored program control and data handling and wireless bidirectional serial communication. A PIC18F4550 is used as the same one embedded in the implant’s prototype. It allows the communication and the collection of data from the implant's MCU in half-duplex mode. Furthermore, it establishes communication as a bridge-link (gateway) between the PC’s GUI interface and the Testbed via USB port.

© SQU-2009 ISSN: 1813-419X

MUSCAT, FEBRUARY 15-18, 2009

185

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09)

Testbed’s MCU 0x00

Interrupt 0x04 (USB_Flag)

Interrupt 0x04 (RX_Flag)

Power-on reset

Read and decode received address

Read digitized data

Enable USB_Flag Interrupt

Send sensor’s address to implant’s MCU

Sleep

Enable RX_Flag Interrupt (USART)

MUSCAT, FEBRUARY 15-18, 2009

fetch task takes roughly 30 seconds at each start-up and directly the program counter (PC) jumps to the user main program.

Compute sensed signal Send analog sensed value to Basestation via USB link RETFIE

Sleep

Figure 5. Cartography of microcontroller’s program memory RETFIE

B. Software program and Monitoring: GUI Interface

Figure 3. Testbed’s routines flow chart

The GUI Interface monitors the acquired signals, receives the data, visualizes and stores them when it is required. The application is developed under LabVIEW platform with the following functionalities:

As we can note, the implant’s MCU has a minimum number of executed instructions in order to reduce power consumption. Further methodology to save power is to minimize the MCU’s running time which will reduce, in consequence, the power consumption. The crucial idea is to refine the firmware implemented into the MCU to perform the implant’s routines flow chart depicted in Fig.4.

ƒ ƒ ƒ ƒ

Communication with Testbed board via USB port Real-time signal acquisition Signal visualization Data storage

Figure 4. Implant’s routines flow chart

Once the firmware is optimized under co-design CAD Tools, its implementation technique is to be intended. Based on bootloader technique, one major advantage of our system is the wireless upgrade memory program of the implant’s MCU (wireless bootloader programming) and the test-bed’s MCU (wired bootloader programming).

Figure 6. Program memory cartography comparison of the firmware implementation methodologies.

V.

A. Evaluation prototypes The overall acquiring chain, Fig.7, is developed for multipurpose use with the following features (some features among them are for future use, such as WiFi and UHF modules):

Fig.6 shows the program memory cartography comparison of the classic technique and the bootloader technique of firmware implementation methodologies. In order for the bootloader to be launched after each reset, a "goto bootloader" instruction must exist within the first 4 instructions. Hence, when the MCU starts-up, it boots to the bootloader code incorporated in the program memory and fetches whether an upgrading firmware is waiting for implementation or not. This

© SQU-2009 ISSN: 1813-419X

EXPERIMENTAL RESULTS AND DISCUSSION

ƒ ƒ

186

Selectable power supply source: Bus power (USB/RS232) or external power. True 5V / 3.3V CMOS compatible drive output and TTL input.

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09)

ƒ

Selectable communication port interface: USB, parallel or serial port (FT232RL/FT245RL). ƒ Full compatible microcontroller chip: PIC16F877, PIC18F4550 and DsPic30F3011 Microchip’ microcontrollers. ƒ Selectable wireless communication data link (UHF, ZigBee or WiFi RF modules) with microcontroller through analogue transmission gate (CD4066). ƒ 8 multiplexed channel sampling with 10 bit resolution (CD4051 1/8 analogue multiplexer). ƒ 10-bit microcontroller enhanced ADC converter. ƒ Programmable gain amplifier with microcontroller through analogue transmission gate (CD4066). ƒ Free Direct DLL driver installation. Drivers for these devices (FT232RL and FT245RL) are provided by FTDI for various operating systems such as Windows and programming languages such as LabVIEW. In this design, a DLL driver for Windows/LabVIEW is used. Since USB provides a single +5 V supply with a current limit up to 500 mA, it is possible to use a USB bus to power the both designed prototypes.

MUSCAT, FEBRUARY 15-18, 2009

2). The signals are recorded during 100 ms and visualized through graphs in real-time.

Figure 8. LabVIEW GUI Interface

In order to evaluate the efficiency of the designed acquiring chain, three different sensors are used to measure temperature, humidity and pressure related to the human body and the ambient air. The calibration of the three sensors is done to get accurate measurements.

The implant’s evaluation prototype was implemented in 9.5cm x 13.5cm, Fig.7. It consists of the programmable analog front end (right side), the PIC18F4550 microcontroller (center) and the ZigBee wireless module (left side). The conditioning block includes: (1) the pressure (strain gauge), humidity and temperature sensors with respectively 0.85 mV/PSI, 40mV/%RH and 10mV/°C as analog output, (2) the analog multiplexer (CD4051 1/8 analogue multiplexer) and (3) the rail-to-rail unipolar low power amplifier (MCP6002).

The pressure sensor is calibrated to get ambient atmospheric pressure which is 1 bar, 14.5037744 PSI (Pounds per Square Inch) or 105 Pascal. As shown in the acquiring graph (fig.8), the pressure is too near the real value with an accuracy of 1 PSI. The temperature sensor is adjusted to get ambient temperature and human body heat with accuracy of 0.5°C related to the sensitivity of the chosen sensor.

The serial to USB interface and bidirectional wireless communication of the Testbed’s evaluation prototype was implemented in 9.5cm x 13.5cm, Fig.7. It consists of: (1) the USB-to-serial UART interface (an USB2.0 full speed compatible FT232RL chip from FTDI Ltd has been chosen because it provides an easy and cost-effective approach to transferring data between peripheral devices and a PC at up to 1 Megabaud [15], (2) the PIC18F4550 microcontroller (center) and (3) the ZigBee wireless module (left side).

The humidity sensor is tuned to get ambient humidity with accuracy of 2 %RH as sensitivity of the chosen humidity sensor. The transmission operating range is measured for different transmitting power level as shown in Fig.9. It’s clear that the operating range is basically limited by the sensitivity of the ZigBee receiver (-92dBm) and the buildings density (urban compactness). Transmitting power level Vs Range 20 18 Operating Range

16 14 12 Transmitting power level Vs Range

10 8 6 4 2

Figure 7. Implant’s and Test-bed’s evaluation board

0 -8

B. Experimental results and discussion

-4

-2

0

Power level (dBm)

Figure 9. Measured operating distance for different power level

Wireless telemetry, USB powering and data transfer, data acquisition and real-time monitoring have been carried out successfully. Fig.8 shows the user interface (man-machine interface) developed under LabVIEW, a visual programming language from National Instruments, allowing easy settings of the Virtual serial port, address sensor and sampling time. The acquired signals are respectively: ambient humidity (Plot 0), ambient pressure (Plot 1) and human body temperature (Plot

© SQU-2009 ISSN: 1813-419X

-6

In the prototype system, a test was conducted to evaluate the bit-error rate (BER) performance of the wireless transmission versus the distance and scaled in transmitting power level. The BER indicates the reliability of the channel. For this purpose, 1 Megabits of data were sent continuously from the implant’s prototype to the test-bed’s prototype. A 187

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP'09) [3]

testing algorithm is implemented into the test-bed’s MCU to check the accuracy of the received bits compared to the randomly sent bits. The BER versus the distance is depicted in Fig. 10. As the distance increases the error rate increases since the transmitter signal is getting weak.

[4]

[5] Error rate Vs distance scaled in transmitting power level 1000000

[6]

900000 Error per 1Mbits

800000 700000

1 mW (0 dBm)

600000

0.631 mW (-2 dBm)

500000

0.398 mW (-4 dBm)

400000

0.251 mW (-6dBm)

300000

0.158 mW (-8 dBm)

[7]

200000 100000

[8]

0 1

10

18.5

24.6

30.8

35.7

40.5

42.8

43

43.8

44.5

Distance (m)

Figure 10. BER measurement Vs distance scaled in transmitting power level

[9]

As reliability of the channel we can choose an operating range around 5 meter which has a low BER starting from 15.2 to 75.78 per 1Megabits respectively to the power level starting from 1mW to 0.158 mW. A 0.251 mW as power level can be affordable with a 65.66 per 1Megabits as bit error rate. The choice of this low power level with such accuracy (low BER) should minimize the power consumption and save the energy.

[10]

[11]

CONCLUSIONS

VI.

[12]

In This paper we have conceived and developed an optimized microcontroller-based architecture for multi-purpose wireless biomedical system based on ZigBee wireless data transceiver. It has the capability to monitor biomedical signals from multiple biosensors by means of different communication standards. Its capability has been tested through standard sensors such as pressure, temperature and humidity sensors with enhanced user graphical interface to visualize and monitor the progress of multi-sensors' curves concurrency in real-time. In addition, we have developed a ZigBee-ready compliant wireless system that offer low power consumption, low cost and advanced network configuration possibilities. Its reliability has been measured and proved through experimental results related to the bit error rate measurement. Hence, an affordable transmitting power level has been chosen in order to reduce the power consumption and save the energy.

[13]

[14]

[15]

This work shows the success of our proof-of-concept study for real-time efficient biomedical evaluation prototype. For this purpose, and in addition to the use of higher transmission frequency for the real implantable system, our future work will be concentrated on a comparative study between UHF, ZigBee and WiFi wireless data link system to evaluate the most adequate wireless link to be compliant with such a biomedical implant. REFERENCES ZigBee Alliance Document 02130, Network Layer Specification, july 2004. [2] M. Hidetoshi and T. Hiroaki, “ZigBee Compliant Sensor Network,” NEC Technical Journal 2006, Vol. 1, no. 1, pp. 102– 105. [1]

© SQU-2009 ISSN: 1813-419X

188

MUSCAT, FEBRUARY 15-18, 2009

OJ. Zhang, J. Zhou, P. Balasundaram, and A. Mason, “A Highly Programmable Sensor Network Interface with Multiple Sensor Readout Circuits,” IEEE Sensors 2003, vol. 2, pp. 748-752. H. Yu and K. Najafi, “Low-Power Interface Circuits for BioImplantable Microsystems,” ISSCC Dig. of Tech. Papers, pp.194-487, Feb. 2003. K. Romer, F. Mattern, and E. Zurich, “The Design Space of Wireless Sensor Networks,” IEEE Wireless Communications, vol.11, no.6, pp.54-61, Dec. 2004. A. Salhieh, J. Weinmann, M. Kochhal, L. Schwiebert, “Power efficient topologies for wireless sensor networks,” International Conference on Parallel Processing, IEEE Computer Society, 2001, pp. 156–166. J. E. Hardy, “Micro Sensor Array Platform,” Fact Sheet, Sensor and Instrument Research Group, OAK Ridge National Laboratory. E. Jovanov, P. Gelabert, R. Adhami, P. Smith, “Real-time portable heart monitoring using low power DSP,” International Conference on Signal Processing Applications and Technology, Dallas, Texas, October 2000. S. L. Toral, J.M. Quero, M.E. Perez, L.G. Franquelo, “A microprocessor based system for ECG telemedicine and telecare,” IEEE Inter. Symp. On Circ. Syst., Syndey, Australia, May 2001, vol. 4, pp. 526–529. W. Bracke, P. Merken, R. Puers, C. Van Hoof “Generic architectures and design methods for autonomous sensors,” Sens. and Actuators A 135 (2007) 881-888. N. Van Helleputte, J.M. Tomasik, W. Galjan, A. Mora-Sanchez, “A flexible system-on-chip (SoC) for biomedical signal acquisition and processing,” (in press), Sens. and Actuators, 2007. Product Manual v1.xAx - 802.15.4 Protocol For OEM RF Module Part Numbers: XB24-...-001, XBP24-...-001 IEEE 802.15.4 OEM RF Modules by MaxStream, Inc. X. Jinwen, C. Yang, A. Mason and P. Zhong “Adaptive MultiSensor Interface System-On-Chip,” IEEE Sensors 2006, EXCO, Daegu, Korea, October 2006. Reid R. Harrison, Paul T. Watkins, Ryan J. Kier, Robert O. Lovejoy, Daniel J. Black, Bradley Greger, and Florian Solzbacher “A Low-Power Integrated Circuit for a Wireless 100-Electrode Neural Recording System,” IEEE Journal Of Solid-State Circuits, vol. 42, no. 1, january 2007. FTDI 2005 FT232RL/FT245RL data sheet and application notes: www.ftdichip.com.