A Wireless Medical Surveillance System

Krause, Michael

A Wireless Medical Surveillance System

Zusammenfassung

Inhaltsangabe:Abstract:
This thesis has the development of a Digital Signal Processor (DSP) based on an Electro-Cardiogram (ECG) analysis system as its main theme. The system measures cardiac signals using two surface ECG leads from which individual heartbeats and pulse trends are extracted. Processed information can be presented on any Bluetooth enabled Personal Digital Assistant (PDA).
The system combines several technologies, e.g. signal measuring and forming unit, DSP hard- and software and a WAP1 server with Bluetooth interface. A basis for this project was a master's thesis that investigates and implements WAP over Bluetooth (see Chapters 5 and 6).
The focus of this work is hardware and software design of the ECG measurement and DSP system. The DSP software includes implementation of medical real-time algorithms for heart beat detection, average beat and pulse trend calculation. All algorithms have been implemented using the C language.

Inhaltsverzeichnis:Inhaltsverzeichnis:
Abstract2
Acknowledgement3
Contents4
Figures6
1.Introduction8
1.1System overview8
1.2Functional description9
2.Introduction into heart anatomy, cardiac signals and measuring methods10
2.1The heart and the atrial contraction10
2.2Cardiac signal characterization and measuring methods12
3.Hardware14
3.1Hardware architecture14
3.2Analog measurement hardware design14
3.2.1System architecture15
3.2.2Patient safety aspects16
3.2.3Detailed system description17
3.3DSP hardware26
3.3.1Development board overview26
3.3.2TI TMS320C5402 DSP features27
3.3.3DSP and DSP board integration29
4.Software31
4.1DSP software architecture31
4.2Main program32
4.3Sensor data acquisition and voltage supervision33
4.3.1McBSP33
4.3.2DMA controller36
4.3.3Timer operation39
4.4Digital signal processing40
4.4.1Signal preprocessing41
4.4.2Beat detection and pulse calculation49
4.4.3Average beat and pulse trend calculation50
4.4.4Preamplifier gain setting51
4.5WAP server communication51
4.5.1UART51
4.5.2EWS commands and UART ISR52
4.6Labview PC-application53
5.Embedded WAP-server55
5.1Wireless Application Protocol55
5.1.1The WAP concept55
5.1.2WAP servers57
5.2Embedded WAP server (EWS)57
5.3Medical surveillance WAP application58
6.Bluetooth interface60
6.1Introduction into Bluetooth60
6.2WAP over Bluetooth (WOB) Implementation62
7.Conclusions63
8.References64
Appendix A - ECG measurements hardware65
Appendix B - DSP software source code73
Appendix C […]

Leseprobe

Inhaltsverzeichnis

ID 7226

Krause, Michael: A Wireless Medical Surveillance System

Hamburg: Diplomica GmbH, 2003

Zugl.: Universität Rostock, Universität, Studienarbeit, 2001

Dieses Werk ist urheberrechtlich geschützt. Die dadurch begründeten Rechte,

insbesondere die der Übersetzung, des Nachdrucks, des Vortrags, der Entnahme von

Abbildungen und Tabellen, der Funksendung, der Mikroverfilmung oder der

Vervielfältigung auf anderen Wegen und der Speicherung in Datenverarbeitungsanlagen,

bleiben, auch bei nur auszugsweiser Verwertung, vorbehalten. Eine Vervielfältigung

dieses Werkes oder von Teilen dieses Werkes ist auch im Einzelfall nur in den Grenzen

der gesetzlichen Bestimmungen des Urheberrechtsgesetzes der Bundesrepublik

Deutschland in der jeweils geltenden Fassung zulässig. Sie ist grundsätzlich

vergütungspflichtig. Zuwiderhandlungen unterliegen den Strafbestimmungen des

Urheberrechtes.

Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in

diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme,

dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei

zu betrachten wären und daher von jedermann benutzt werden dürften.

Die Informationen in diesem Werk wurden mit Sorgfalt erarbeitet. Dennoch können

Fehler nicht vollständig ausgeschlossen werden, und die Diplomarbeiten Agentur, die

Autoren oder Übersetzer übernehmen keine juristische Verantwortung oder irgendeine

Haftung für evtl. verbliebene fehlerhafte Angaben und deren Folgen.

Diplomica GmbH

http://www.diplom.de, Hamburg 2003

Printed in Germany

Lund University, Lund Institute of Technology, Department of Electroscience

Abstract

This thesis has the development of a Digital Signal Processor (DSP) based Electro

Cardiogram (ECG) analysis system as its main theme. The system measures cardiac signals

using two surface ECG leads from which individual heartbeats and pulse trends are extracted.

Processed information can be presented on any Bluetooth enabled Personal Digital Assistant

(PDA).

The system combines several technologies, e.g. signal measuring and forming unit, DSP hard-

and software and a WAP

server with Bluetooth interface. A basis for this project was a

master's thesis that investigates and implements WAP over Bluetooth (see Chapters 5 and 6).

The focus of this work is hardware and software design of the ECG measurement and DSP

system. The DSP software includes implementation of medical real-time algorithms for heart

beat detection, average beat and pulse trend calculation. All algorithms have been

implemented using the C language.

WAP Wireless Application Protocol

Lund University, Lund Institute of Technology, Department of Electroscience

Acknowledgement

I want to thank my supervisor and friend Martin Stridh at LTH Lund for his help, support and

all the time that he has spent to answer my questions. He had always time to listen to me and

gave me many helpful hints.

Further, I am very thankful to Professor Leif Sörnmo, LTH Lund and Professor Lienhard

Pagel, University of Rostock for supporting me. They made the project performed at LTH

Lund possible.

People who have mainly contributed to this project are Tomas Mandorf and Philippe Burlion

at AU-System, Lund. They have developed the WAP-server and the WAP over Bluetooth

application, which are important parts of the surveillance system.

Further, thanks to the staff at the Department of Electroscience for helping me in the daily

work and producing the circuit board.

Finally, I am grateful to my family and Andrea for their love and patience during my work in

Lund.

Lund University, Lund Institute of Technology, Department of Electroscience

Contents

Abstract ... 2

Acknowledgement... 3

Contents... 4

Figures... 6

1. Introduction ... 8

1.1 System

overview ... 8

1.2 Functional

description ... 9

2. Introduction into heart anatomy, cardiac signals and measuring methods ... 10

2.1

The heart and the atrial contraction... 10

2.2

Cardiac signal characterization and measuring methods ... 12

3. Hardware ... 14

3.1 Hardware

architecture ... 14

3.2

Analog measurement hardware design ... 14

3.2.1 System

architecture ... 15

3.2.2 Patient safety aspects... 16

3.2.3 Detailed system description ... 17

3.3 DSP

hardware... 26

3.3.1 Development board overview ... 26

3.3.2 TI TMS320C5402 DSP features ... 27

3.3.3 DSP and DSP board integration ... 29

4. Software ... 31

4.1

DSP software architecture... 31

4.2 Main

program... 32

4.3

Sensor data acquisition and voltage supervision... 33

4.3.1 McBSP ... 33

4.3.2 DMA

controller ... 36

4.3.3 Timer

operation ... 39

4.4

Digital signal processing ... 40

4.4.1 Signal

preprocessing ... 41

4.4.2 Beat detection and pulse calculation ... 49

4.4.3 Average beat and pulse trend calculation... 50

4.4.4 Preamplifier gain setting ... 51

Lund University, Lund Institute of Technology, Department of Electroscience

4.5

WAP server communication ... 51

4.5.1 UART ... 51

4.5.2 EWS commands and UART ISR ... 52

4.6 Labview

PC-application... 53

5. Embedded

WAP-server... 55

5.1

Wireless Application Protocol ... 55

5.1.1 The

WAP

concept ... 55

5.1.2 WAP

servers... 57

5.2

Embedded WAP server (EWS) ... 57

5.3

Medical surveillance WAP application... 58

6. Bluetooth

interface ... 60

6.1

Introduction into Bluetooth ... 60

6.2

WAP over Bluetooth (WOB) Implementation... 62

7. Conclusions ... 63

8. References ... 64

Appendix A ECG measurements hardware ... 65

Appendix B DSP software source code ... 73

Appendix C Labview PC application source code... 90

Lund University, Lund Institute of Technology, Department of Electroscience

Figures

Figure 1.1-1: Block diagram of the ECG system... 8

Figure 2.1-1: The anatomy of the human heart... 10

Figure 2.1-2: The conduction system of the heart... 11

Figure 2.2-1: Typical electrocardiogram... 12

Figure 3.1-1: Block diagram of the signal measurement and processing units ... 14

Figure 3.2.1-1: Block diagram of the signal measuring hardware... 16

Figure 3.2.3-1: Schematic view of a typical instrumentation amplifier... 17

Figure 3.2.3-2: Block diagram of the PGA204 instrumentation amplifier ... 18

Figure 3.2.3-3: Principle of optocouplers ... 18

Figure 3.2.3-4: Optocouplers sub-circuit ... 19

Figure 3.2.3-5: Preamplifier gain settings... 19

Figure 3.2.3-6: Burr-Brown ISO 122 block diagram... 20

Figure 3.2.3-7: Aliasing in frequency range ... 21

Figure 3.2.3-8: 4

order Butterworth low-pass filter ... 22

Figure 3.2.3-9: Impulse response of the 4

order antialiasing filter ... 24

Figure 3.2.3-10: ECG amplifier and level shift sub-circuit ... 24

Figure 3.3.1-1: DSP development board overview... 27

Figure 3.3.2-1: TI'5402 properties... 28

Figure 3.3.2-2: TI'5402 memory map... 29

Figure 4.2-1: Main program structure... 32

Figure 4.3.1-1: McBSP general structure... 34

Figure 4.3.1-2: SPI interface ... 35

Figure 4.3.1-3: SPI mode pin overview ... 35

Figure 4.3.1-4: MAX 186 control byte format... 36

Figure 4.3.2-1: DMA register subaddressing... 37

Figure 4.4-1: Digital signal processing block diagram ... 40

Figure 4.4.1-1: Preprocessing stages of the ECG signal... 42

Figure 4.4.1-2: FIR filter structure... 43

Figure 4.4.1-3: Digital filter design using the windowing method ... 44

Figure 4.4.1-4: Ideal low-pass filter and filter length limitation... 45

Figure 4.4.1-5: Preprocessing low-pass filter design ... 46

Lund University, Lund Institute of Technology, Department of Electroscience

Figure 4.4.1-6: Filter characteristic and impulse response of low-pass, band-pass and

smoothing filter ... 47

Figure 4.4.1-7: Impulse response of a Hilbert transformer ... 48

Figure 4.5.1-1: UART working principle ... 52

Figure 4.5.2-1: EWS commands overview ... 52

Figure 4.6-1: Labview PC-application... 53

Figure 5.1.1-1: WAP stack compared to the OSI stack model ... 56

Figure 5.3-1: WAP application start page... 58

Figure 5.3-2: Measurement selection page ... 58

Figure 5.3-3: Heartbeat mode page... 59

Figure 6.1-1: Bluetooth stack... 61

Lund University, Lund Institute of Technology, Department of Electroscience

1. Introduction

Many people suffer from some heart problems that make it difficult to move around freely.

Common ECG systems expect the patient to be in a hospital connected to dozens of wires.

The wireless medical surveillance system for ECG measurements and analysis is a new

application, which may offer at least some patients more freedom. For example the wireless

system can be used to record a 24h ECG or by including a mobile phone to the system a

physician can check the patients heart from somewhere else. The technical novelty of the

system is the combination of several new technologies. The surveillance system is the first

application that combines a DSP with medical real-time algorithms, an embedded WAP

server (EWS) and a Bluetooth interface to a whole system. Not all parts of the system have

been developed in this thesis. An overview of the system is given in Section 1.1 and a brief

description of how the different parts work together is found in Section 1.2.

1.1 System

overview

The system consists of two main parts: patient and presentation unit. The patient unit collects

data by ECG sensors, processes information and communicates via a wireless interface with

the presentation unit. A Bluetooth enabled PDA, mobile phone or other devices with a WAP

browser can be used as a presentation unit. Therefore the system is very flexible and covers a

wide range of applications. This thesis is mainly focused on the patient unit. Figure 1.1-1

shows a block diagram of the system.

Figure 1.1-1: Block diagram of the ECG system

As shown above, the patient unit includes four sub-units:

ECG measurements unit

The ECG measurements unit measures cardiac information, amplifies the analog signal and

converts it into digital data. Only two surface ECG sensors are needed to measure cardiac

activity. Furthermore the hardware must fulfill some patient safety requirements to avoid

hazards in case of system malfunctions.

Lund University, Lund Institute of Technology, Department of Electroscience

DSP unit

A fast DSP performs data communication and medical real-time analysis. Purposes of the

DSP software are real-time sensor data acquisition, medical calculations and WAP server

communication. A basic element for medical analysis is the beat detection algorithm, which

allows finding of heart beats in recorded sensor information.

WAP server

The WAP server handles communication between Presentation and DSP unit. In general a

WAP server provides Internet services to wireless networks. In this application, an embedded

WAP server (EWS) developed by AU-System, Lund offers ECG analysis via a Bluetooth

interface.

Bluetooth interface

Wireless communication needs a well-defined and safe interface. Therefore the Bluetooth

standard is used to offer high-speed wireless data transfer.

Development of measuring and DSP sub-units are part of this project. An implementation of

the embedded WAP Server including a Bluetooth interface was covered in another master's

thesis [1], performed at AU-System, Lund.

1.2 Functional

description

Since the WAP over Bluetooth communication is used, the user needs a Bluetooth enabled

Presentation unit with WAP communication ability for this ECG application. Nowadays most

devices like mobile phones or PDAs have these possibilities. In the first step the WAP

browser in the PDA searches and contacts the EWS over its Bluetooth interface. When an

EWS was found and a connection could be installed, the PDA WAP Browser reads the ECG

application WML

page on the server and presents it on the PDA screen. On this page the user

defines patient preferences such as name, weight and body size. In the next step measuring

methods and sample rate are selected. The surveillance system offers four different measuring

modes: electrocardiogram, pulse rate, pulse trend and average beat. All input data is sent to

the EWS. A CGI script

is executed on the EWS managing data communication between the

EWS and the DSP. In this way input data like measuring mode and sample rate is sent

forward to the DSP. The DSP has real-time based software for ECG sensor data acquisition

and medical analysis. For statistical calculations, e.g. pulse trend or average beat calculation,

sensor data acquisition and medical analysis algorithms must be executed continuously. After

receiving a request from the EWS, the DSP sends required information to the EWS where a

new WML page is created. The created WML page includes a picture of requested

information, which is transmitted over the Bluetooth interface and presented on the PDA.

This WML page can then be arbitrarily updated.

WML

Wireless Markup Language, CGI script

Common Gateway Interface script

Lund University, Lund Institute of Technology, Department of Electroscience

Introduction into heart anatomy, cardiac signals and

measuring methods

To be able to design a system for ECG measurements and analysis it is important to know

some basic facts about the human heart, its working principle and how cardiac activity can be

measured. This knowledge is needed to understand ECG measuring method, hardware design

and medical algorithms described in the following chapters. A brief introduction of the heart

anatomy and the working principle of the heart is given in Section 2.1. Section 2.2 describes

measuring methods for cardiac activity and characterizes the measured signals.

2.1 The heart and the atrial contraction

The heart is known as the engine of life. It organizes the blood flow and allows the body to be

supplied with oxygenated blood from the lungs. Thus, the heart is an essential organ for life.

The human heart is divided into two halves. Each half includes an atrium and a ventricle.

Between the atria and the ventricles is a wall with atrioventricular valves. These valves open

and close by pressure variations due to blood flow. They are open during an atria contraction

and closed when the ventricle contracts. This allows the blood to flow from the atrium to the

ventricle without any reverse flow. Two other kinds of valves control the flow from and to the

heart. They are situated between the ventricular outflow tracts and the arteries, i.e. between

pulmonary artery and aorta. All valves work in the same way. They are pushed open and

forced to close by blood pressure. Figure 2.1-1 shows the anatomy of the human heart.

Figure 2.1-1: The anatomy of the human heart

The heart consists of tightly bound muscle cells that enclose the blood in the heart. In order to

beat all the muscle cells must contract almost simultaneously. Therefore a fast activation of

all cells is needed. An electrical activation is optimal because the transmission of the

activation can be done very rapidly. Inside the heart there are some cells that are specialized

for fast electrical transmission. These cells form a conduction system; see Figure 2.1-2.

Lund University, Lund Institute of Technology, Department of Electroscience

Figure 2.1-2: The conduction system of the heart

All cells of the conducting system have the property to increase its membrane potential

spontaneously. The increase is slow until a critical threshold is reached. If the critical level is

reached the membrane potential increases very rapidly to a certain level. This process is called

depolarization. Between 100 and 300ms later the membrane potential decreases to its original

level (repolarization). The time before the potential is restored depends on the place where the

cell is situated. During this depolarization the cell cannot be reexcited by an electrical

activation. The process of depolarization and repolarization has automaticity, which means

that the cells of the conducting system are activated periodically without external activation,

e.g. by the autonomic nervous system (ANS). Different cells of the conducting system have

automaticity with different intrinsic rates. A cardiac cycle starts in the sinoatrial (SA) node.

The SA node is situated in the upper part of the right atrium and has an intrinsic rate between

50 and 100 per minute. The atrioventricular (AV) node in the wall of right atrium and

ventricle has a rate of 40-45 per minute. This node links the electrical impulse from the atria

to the ventricles. The bundle of His and the Purkinje fibers that spread the impulse into the

ventricles have intrinsic rates less than 35 per minute. Because of the fastest rate the SA node

normally determines the beat rate and therefore its cells are often called pacemaker cells. If

the SA node fails to initialize heartbeats some other parts of the conducting system with lower

intrinsic rate will take over. Beats that are not initiated in the SA node are called ectopic

beats. After an impulse is generated in the SA node the electrical potential propagates from

cell to cell throughout the heart. First the impulse spreads around the SA node in the right

atrium and then it goes on to the left atrium. The impulse is so fast that left and right atrium

are depolarized at almost the same time. During the resulting contraction the ventricles will be

filled with blood. In the next step the electrical impulse reaches the AV node and

depolarization comes to the ventricles. The impulse propagation in the AV node is very slow

and causes a delay of about 0.1s. Because of the impulse delay the blood can flow from the

atrium to the ventricles before the ventricular contraction starts. The electrical activation

propagates from the AV node through the bundle of His that is divided into left and right

bundle branches to the Purkinje fibers. These fibers consist of large conducting cells that

quickly spread the impulse throughout the ventricles. Therefore the ventricles depolarize at

approximately the same time and the following contraction from the bottom of the heart

upwards forces the blood to flow into the aorta and the pulmonary artery.

Lund University, Lund Institute of Technology, Department of Electroscience

Although the beat process has automaticity and does not require active control, the autonomic

nervous system (ANS) is able to affect the force and the frequency of cardiac contraction. For

this purpose the ANS has two subsystems, the sympathetic and parasympathetic nervous

system. The sympathetic system is excitatory what means it can increase the firing rate of the

SA node and affects the conduction velocity through the Purkinje fibers. The parasympathetic

system has an opposite effect on the heart. It decreases SA node firing rate and conduction

velocity. In this way the ANS can control the cardiac activity and adapts beat frequency and

force optimally to the physical conditions of the body.

2.2 Cardiac signal characterization and measuring methods

Cardiac signals are usually measured using the electrocardiogram (ECG) method. Several

electrodes on the body surface measure electrical activity from the heart. Beat information is

recorded and shown in a diagram called electrocardiogram. The ECG is a standard tool in

medical clinics for diagnosis of cardiac diseases, e.g. atrial fibrillation.

The ECG measures the field around the propagated electrical impulse in the heart. When a

depolarization wavefront moves towards an ECG sensor a positive potential can be measured

at the electrode. A negative potential is measured while the wavefront is moving away. In the

other case during the repolarization the sensor potential is negative when the wavefront

moves towards the sensor and the potential is positive while moving away. For spatial

information the ECG is measured by several sensors. Usually there are three different ECG

lead systems for measuring:

- Three bipolar leads on left arm, right arm and left leg

- Six precordial leads

- Three augmented unipolar leads

All leads together form the 12-lead standard ECG. For the medical surveillance system

developed in this thesis only the three leads on arms and left leg are used. Therefore the

system can be used in a mobile environment but the ECG quality is decreased compared to a

12-lead standard ECG. In spite of using only three electrodes the signal quality must be able

to reflect patients heart activity; see Figure 2.2-1.

Figure 2.2-1: Typical electrocardiogram

Lund University, Lund Institute of Technology, Department of Electroscience

Figure 2.2-1 shows an electrocardiogram of a normal subject. A cardiac cycle starts with a P-

wave. This wave is a result of atrial depolarization. The following QRS complex occurs

during the ventricular depolarization phase. Afterwards the T-wave reflects ventricular

repolarization. Atrial repolarization occurs at the same time as the QRS complex occurs.

Thus, atrial repolarization is invisible in the ECG.

Lund University, Lund Institute of Technology, Department of Electroscience

3. Hardware

The hardware of the medical surveillance system consists of several parts: The analog

hardware including ECG electrodes, the DSP hardware, the hardware for the EWS including a

Bluetooth interface and the PDA. This chapter explains the parts that have been developed in

this thesis. After an overview of the hardware on which this thesis is focused on (see Section

3.1) the analog measuring hardware is explained in more detail in Section 3.2. Finally Section

3.3 deals with the DSP hardware that has been used for implementation of medical real-time

algorithms.

3.1 Hardware

architecture

As mentioned before, this thesis is focused on the first two parts of the patient unit (compare

figure 1.1-1). The sub-units include analog ECG signal measuring, signal adaptation, analog

to digital conversion and digital signal processing. Figure 3.1-1 shows a block diagram of the

signal measurement and processing units.

Figure 3.1-1: Block diagram of the signal measurement and processing units

Two ECG electrodes measure the potential difference between patient left and right arm.

Under normal conditions the analog signal is similar to the signal shown in figure 2.2-1.

Because of low signal amplitudes (approximately 0.5-2mV) the signal must be amplified and

adapted to the analog to digital converter (ADC) input range in the analog measurement unit.

Afterwards a fast ADC converts the analog signal into its digital complement. The DSP

records ECG sensor data and the software analyzes the cardiac signals in real-time.

Communication between embedded WAP server and DSP is done via a RS232 interface

including a UART

3.2 Analog measurement hardware design

Analog measurement hardware has several functions. Its main function is to acquire beat

information, amplify electrode signals and convert analog data into digital information. But

there are some more requirements to the hardware concerning signal quality and patient

UART Universal Asynchronous Receiver and Transmitter

Lund University, Lund Institute of Technology, Department of Electroscience

safety. Patient safety is the most important fact in all kinds of medical applications. In case of

malfunction the patient must be save and any contact to dangerous electrical potential must be

avoided. The following sections describe the analog hardware design in order to fulfill all

requests and how the hardware takes care of patient safety.

3.2.1 System

architecture

Two electrodes, one on each arm of a patient measure cardiac activity. An instrumentation

amplifier increases the voltage difference between the electrodes to an amplitude level of

approximately

±0.5V. The voltage amplitude measured by the electrodes depends on different

parameters, i.e. skin and arm thickness or physical condition. Therefore the instrumentation

amplifier has a programmable gain of 1, 10, 100 or 1000 (see Section 3.2.3), which allows the

use of the same ECG sensors for different patients. The third ECG electrode is attached to

patient leg or neck and connects patient and measuring ground potential. Since the real

ground can be floating the electrode potential is controlled by a control circuit that limits the

input common mode voltage of the instrumentation amplifier to a certain level. After the

preamplification the ECG signal crosses an isolation barrier using an isolation amplifier.

Isolation amplifiers offer galvanic separation in a circuit, which is recommended to fulfill

patient safety aspects (compare Section 3.2.2). Theoretically the output signal can be directly

converted into digital data for DSP calculations. Due to data sampling in digital signal

processing, a direct conversion would cause some problems such as aliasing effects. Aliasing

occurs when the sampling rate is to low compared to the signal frequency range. To be sure

that there is no aliasing, a 4

order low pass filter has been inserted in order to limit signal

frequency and fulfill sampling theorem conditions. The following operational amplifier

adjusts the bi-directional signal (

±0.5V) to a unidirectional signal between 0 and 4V. Now, the

signal adaptation is finished and the analog ECG signal can be converted into digital data. The

MAX 186 analog to digital converter, which converts data, has 12bit resolution and a serial

port interface for data transfers to the DSP. Figure 3.2.1-1 shows a block diagram of the

signal measuring hardware. The detailed circuit is found in Appendix A.

Lund University, Lund Institute of Technology, Department of Electroscience

Figure 3.2.1-1: Block diagram of the signal measuring hardware

3.2.2 Patient safety aspects

As stated before, patient safety must be guaranteed in all possible hardware states. Beside

normal states there can be abnormal states caused by defect circuit parts or overvoltages in the

power supply. For this reason ECG measurement hardware has some extra precautions to

protect patients against hardware malfunctions and avoid dangerous hazards. First, there is a

galvanic separation of the measuring hardware into patient side and signal adaptation and

processing side. A high isolation between both circuits makes it almost impossible to get in

direct contact with any dangerous voltage from the unisolated side. The unisolated side is the

part where analog filter, amplifier, ADC and DSP are. Because of the isolation it is necessary

to find a way to across the galvanic barrier for some signals and power supply in the isolated

part. Isolation amplifier and optocouplers allow crossing of the isolation for analog and digital

signals (see Figure 3.2.1-1). A special DC/DC converter IC is used to power the isolated part

from the unisolated side. Additionally to galvanic separation, high resistance resistors

(300kOhms) limit any possible current in the ECG electrode wires. The resulting body current

is the physical phenomenon, which can be dangerous when someone is in contact to electrical

voltages.

All above mentioned precautions help to make the surveillance system safe and protect

patients optimally against any hazards.

Lund University, Lund Institute of Technology, Department of Electroscience

3.2.3 Detailed system description

The development of measuring hardware contains several steps. After recognizing all

requirements a general structure (compare Figure 3.2.1-1) has been designed. Every part was

constructed independently and afterwards the circuit board has been created physically by

using the EAGLE CAD system. This sub-section explains the circuit development process

beginning at the ECG signal origin.

Instrumentation amplifier

An instrumentation amplifier is a special type of operational amplifier. It differs from normal

operational amplifiers by that it includes integrated amplifiers with integrated feedback

resistors to perform a subtraction of the input signals. For ECG purposes, the instrumentation

amplifier subtracts the input voltage between left and right arm electrodes and amplifies the

voltage difference. A schematic view of a typical instrumentation amplifier is shown in Figure

3.2.3-1.

Figure 3.2.3-1: Schematic view of a typical instrumentation amplifier

The integrated amplifiers (A1...A3) are similar to common operational amplifiers. They are

based on the principle of differential amplifiers. Two external resistors set the gain of

instrumentation amplifiers. Equation (1) describes the relation between input and output

voltage:

)

)(

(

(1)

For ECG measurements variable gain setting is necessary to adapt different sensor voltage

levels, caused by different patients, to the hardware. Adjustment can be done either analog or

digital. Analog adjustment means that an analog control circuit changes the resistor value.

However, digital adjustment changes the resistance by setting some bits. Analog resistor

setting has some advantages especially since the resolution is unlimited. Nevertheless digital

gain setting is easy to use and to implement. Some unused DSP pins can set the gain of the

instrumentation amplifier. Therefore a digital programmable instrumentation amplifier (Burr-

Brown PGA204) has been used in this application. Two bits set the gain to a value of 1, 10,

100 or 1000 to cover a wide input voltage range. Figure 3.2.3-2 shows a block diagram of the

PGA204 instrumentation amplifier.

Lund University, Lund Institute of Technology, Department of Electroscience

Figure 3.2.3-2: Block diagram of the PGA204 instrumentation amplifier

The gain setting is controlled by software running on the DSP. Section 4.4.4 describes the

strategy how the DSP software controls the amplifier gain. Two digital DSP output bits set the

gain physically. Because of the patient the safety requirements in medical systems, DSP and

preamplifier are situated in different circuit parts, which are isolated against each other. For

gain setting the electrical barrier is crossed using optocouplers.

Optocoupler

Optocouplers are optical-electrical devices that are used in a wide range of applications, e.g.

when a galvanic separation must be passed. Crossing an isolation barrier is normally done

using some kind of modulation and a following demodulation. Optocouplers are based on

light modulation and demodulation, which means that the electrical current is transformed

into light and later the light is retransformed into current. The working principle is as follows:

A light emission diode (LED) converts the electrical current into light. Then, the light passes

the isolation and hits the basis of a bipolar phototransistor where the light influences the

transistor basis potential. Thereby the collector current can be set through the exposure of

light. The LED works as the modulator and the phototransistor is the demodulator; see Figure

3.2.3-3.

Figure 3.2.3-3: Principle of optocouplers

Lund University, Lund Institute of Technology, Department of Electroscience

Optocouplers are able to transmit both, analog and digital signals although the transmission of

analog signals causes some difficulties. First, the transmission characteristic is nonlinear and

often a linearization is needed e.g. by adding an offset to the signal. Secondly, optocouplers

invert the signal and the transmission is unidirectional. For bi-directional communication two

optocouplers are necessary. Since the gain setting of the preamplifier is digital, optocouplers

only cause an inversion in the digital logic. This signal inversion must be considered when the

DSP software sets the gain. Figure 3.2.3-4 shows the gain setting sub-circuit. The ECG

hardware needs two optocouplers, one to set each bit.

Figure 3.2.3-4: Optocouplers sub-circuit

As shown in the figure above, the HCPL2530 type optocouplers from Hewlett Packard are

used. This type provides digital data transmission and is compatible to the TTL and CMOS

signal standard. Furthermore a manual gain setting for testing and developing purposes has

been incorporated in the design. Therefore two jumpers have been added to the circuit; see

Figures 3.2.3-4 and 3.2.3-5.

Jumper 0

Jumper 1

Gain setting

open open

1000

1-2 open

100

open 1-2

1-2 1-2

1-2 2-3

not

allowed

2-3 1-2

not

allowed

2-3 open

not

allowed

open 2-3

not

allowed

2-3 2-3 software

programmable

Figure 3.2.3-5: Preamplifier gain settings

Lund University, Lund Institute of Technology, Department of Electroscience

Isolation amplifier

Because of the galvanic separation the measured and preamplified analog signal must cross

the isolation barrier. There are several possibilities to transmit an analog signal over an

electrical isolation:

- Optocouplers for analog data transmission

- DC/DC converter

- Isolation amplifier

As described earlier, optocouplers for analog data transmission can cause some linearity

problems and common DC/DC converters often have low isolation or influence the signal

characteristic. In medical environment the isolation should separate continues voltages higher

than 1000Vrms (AC, 60Hz). The requirements for voltage peaks are higher than for continues

voltages, e.g. 2400V for 1s. Special isolation amplifiers can provide such high isolations.

Isolation amplifiers use a galvanically separated input and output section, which is isolated by

isolating capacitors. The matching capacitors are very small (ca. 1pF) but their break through

voltage is very high. In order to transmit voltages with low frequency (down to 0 Hz) a

modulation is needed to across the isolation. Therefore the input and output sections include a

modulator and demodulator. The input signal is duty-cycle modulated and transmitted

digitally across the barrier. The output section receives the signal and reconverts it into analog

data. In this application the ECG measurement hardware uses the Burr-Brown ISO 122

isolation amplifier. Beside high isolation (1500Vrms, 60Hz, continuously) this type provides

high precision (0.016% max. nonlinearity, max. 0.5 %FSR

gain error) and wide supply

power range from

±4.5V up to ±18V. Figure 3.2.3-6 shows the ISO 122 block diagram.

Figure 3.2.3-6: Burr-Brown ISO 122 block diagram

The modulation starts with an integrator circuit (A1 in Figure 3.2.3-6), which integrates the

difference between the input current (V

/200kOhm) and a fixed switchable

±100µA reference

current. The integrator current loads the 150pF capacitor until the comparator threshold is

reached. Then, comparator and sense amplifier force the current source to switch. If V

=0V

the result is a triangular waveform with 50% duty cycle. The internal oscillator forces the

Details

Seiten
Erscheinungsform: Originalausgabe
Jahr: 2001
ISBN (eBook): 9783832472269
ISBN (Paperback): 9783838672267
DOI: 10.3239/9783832472269
Dateigröße: 3.9 MB
Sprache: Englisch
Institution / Hochschule: Universität Rostock – Elektronik und Informationstechnik
Erscheinungsdatum: 2003 (September)
Note: 1,0
Schlagworte: programmierung bluetooth patientenüberwachung

Autor

Michael Krause (Autor:in)

A Wireless Medical Surveillance System

Zusammenfassung

Leseprobe

Inhaltsverzeichnis

Details

Autor

Michael Krause (Autor:in)