Introduction:
Epilepsy is a very fatal condition which is caused as a result of imbalance in the nervous
system. The very common symptoms of epilepsy includes sudden fluctuations in heart beat rate
and involuntary muscular movements (seizures). The aura (practical symptom) of epilepsy
includes fluctuations
in heartbeat, nausea, dizziness etc.
The wireless electronic diagnosing system designed here is exclusively meant for epilepsy
patients. The system helps them in accurately predicting the occurrence of seizures. Sudden
occurrence of seizures during driving may lead to accidents and its occurrence during sleeping
hours can even
lead to the patient’s death, if no immediate, proper attention is provided by a bystander or a
doctor. With the aid of this system, the patient can lead a normal life. Since the occurrence of
seizures is unpredictable, it will be a very risky task to leave the patient alone.
The electronic system presented here is a wearable device which predicts the occurrence of
epilepsy in a few minutes advance. The device utilizes the signals from human body to detect the
occurrence of epilepsy. As soon as the device detects the symptoms, it transmits a coded signal.
The signal
is decoded by a wireless receiver to produce control signals for switching an alarm device,
mobile messaging device and an automatic vehicle control system appropriately. In future, GPS
could be incorporated to trace out the exact location of the patient.
Current technologies for acquiring signals from the patient’s body are very much developed.
Many sensors are available which can detect the heart beat and muscular movements noninvasively
and accurately.
Such non invasive technique for measuring heart beat is pulse oximetry. Using this technique,
heart beat can be accurately monitored. Muscular convulsions are collected using micro
electromechanical sensors (MEMS) firmly attached to the body. The sensors used are small in
size and can be firmly attached to the body. The accelerations resulting from epileptic
convulsions are sensed using MEMS accelerometer which is very accurate, precise and small in
size. To provide wireless communication channel low cost network using DATEx protocol is
utilized. DATEx is a standard protocol developed by National Instruments.
Heart beats are to be monitored continuously. Any sudden variation in heart beat which is caused
by the onset of epileptic seizures is detected and confirmed with the MEMS signal. When the
seizure is confirmed, message is transmitted to the surroundings for initiating necessary
protective measures for the patient.
The device is designed as a wireless, wearable and personal equipment. The device can sense the
aura of pre ictal stage in a few minutes advance and takes the necessary safety measures
automatically. Hence a technician’s assistance is not required for the patient. Therefore this
device will be extremely useful for patients (especially youngsters) who wish to be active in their
life. The user gets absolute freedom from wires and can be used when moving.
To practically implement the epilepsy prediction system, the following aspects should be
implemented.
1. Sensing biometric signals: Two types of biological signals are required for processing. They
are heart beat and muscular convulsions. The heart beat be measured using PCI DIO 32 HS and
mascular movements can be measured by an accelerometer.
2. Processing it and taking decisions: Processing of the signals is done by software programmed
into a microcontroller. The software is designed in such a way that it detects the exact symptom
of epilepsy.
3. Communication: Communication is set up using a transmitter and receiver module with
DATEx protocol.
4. Controlling: Automatic vehicle control system, mobile messaging device and an alarm device
PXI-6608 is integrated to the receiver for protecting the patient.
Constraints
1. Smaller size and weight requirement
2. Low power consumption requirement as the device is battery operated.
3. Suitable long life battery
4. Accurate technique or algorithm for foolproof detection of seizure
5. Secure communication between the wearable equipment and the receiving unit
6. Signal processing requirement
7. Cost effectiveness
The application of this system focus on epilepsy patients who wish to move freely without the
assistance of others in their life. The system is a wearable device which can detect the aura of
seizure in an epilepsy victim very much precisely in time by processing the signals available
from the patient’s body at the predictal state. The system uses a processing device to process the
signals from the human body and activates a wireless transmitter which transmits a coded signal.
The receiver decodes the signal using another processing unit which results in the production of
control signals for activating various safety devices mounted on the vehicle or on the dormitory
where the patient commonly resides. For e.g, if the seizure occurs while the victim driving a
vehicle or while sleeping the device automatically generates
control signals for the control of vehicle, setting off an alarm circuit and messaging the doctor
about the patient’s condition via short messaging Service (SMS). The system can be expanded
easily in such a way to include Global Positioning System (GPS) for tracing out the exact
position of the victim of epilepsy in the future. Thus the device saves the patient from accident or
even death and acts as a “LIFE SAVER”.
Design
The design consists of hardware and software sections. The device hardware mainly consist of
three parts namely, (i) Heart beat sensor, (ii) Seizure detector, (iii) Processor and (iv) Wireless
transceiver
(i) Heart beat sensor: The heart beat of the patient is to be monitored. For this purpose, a PCI
DIO 32 HS is used. PCI DIO 32 HS measures heart beat by sensing the difference in absorbance
of infrared radiation by blood during systolic and diastolic activities of heart. The volume of
blood flowing through arteries varies widely during each heart beat. Hence if infrared radiation is
incident on it, the absorbance of IR also varies according to the heart beat. These variations are
sensed using a photo detector to determine the heart
beat.
The PCI DIO 32 HS designed here works using reflective principle. The IR source emits IR
radiation which is reflected in accordance with the flow of blood. The reflected rays are detected
using a photo detector.
A sensor is placed on a thin part of the patient's anatomy, usually a fingertip or earlobe, and light
of infrared wavelength is made incident on the body. Changing absorbance of the infrared is
measured, allowing determination of the absorbance’s due to the pulsing arterial blood alone,
excluding venous blood, skin, bone, muscle, fat, and (in most cases) fingernail polish. The circuit
of PCI DIO 32 HS consists of a trans-resistance amplifier, voltage follower, difference amplifier,
and filter. All these stages are cascaded together to from the complete circuit of PCI DIO 32 HS.
The circuit works in 5 V supply. In order to get perfect amplification sans noise, ultra low offset
operational amplifier OP07 and FET input operational amplifier LF 356N is selected. A trans
resistance amplifier is used in the first stage to convert the photodiode current to voltage. The
major design of this sensor is its output voltage and the output frequency. The output frequency
is band limited to 15 Hz using filters. Low pass first order butterworth filter is used. Low pass
filter is designed at 15 Hz upper cut off frequency with a gain of 1.5. A high pass first order
butter worth filter with lower cut off frequency of 0.5 Hz is cascaded with the low pass to
remove the dc voltage. An amplifier is set at the output of the PCI DIO 32 HS in order to raise
the output signal level to +5V (approx). Amplifier with amplification factor of 50 is designed.
Typical output of the sensor is shown on the graph below. Normal heart beat is 72 beats per
minute. That is the frequency of the signal is 1.2 Hz for a healthy person. The output amplitude
varies from 70mV to 120mV.
(ii) Seizure detector: Seizures are involuntary muscular movements which occur during
epilepsy. Muscular movements are sensed using MEMS (micro electro mechanical sensor)
accelerometer. A 3D accelerometer is used to sense the muscular movements. The SCXI-1530 is
a low cost, low power, complete 3-axis accelerometer with signal conditioned voltage outputs,
which is all on a single monolithic IC. The SCXI-1530 is a complete acceleration measurement
system on a single monolithic IC. The SCXI-1530 has a measurement range of ±3 g. The sensor
is a polysilicon surface-micro
machined structure built on top of a silicon wafer. Polysilicon springs suspend the structure over
the surface of the wafer and provide a resistance against acceleration forces. Deflection of the
structure is measured using a differential capacitor that consists of independent fixed plates and
plates
attached to the moving mass. The fixed plates are driven by 180° out-of-phase square waves.
Acceleration deflects the beam and unbalances the differential capacitor, resulting in an output
square wave whose amplitude is proportional to acceleration. Phase-sensitive demodulation
techniques are then used to rectify the signal and determine the direction of the acceleration.
The demodulator’s output is amplified and brought off-chip through a 32 k resistor. The signal
bandwidth of the device is set by adding a capacitor. This filtering improves measurement
resolution and helps prevent aliasing.
Performance of the project was affected due to the non availability of 3 axis accelerometer.
Hence here I have used a single axis accelerometer MMA1260EG from FREESCALE
semiconductor for detecting the muscular convulsions. It has a sensitivity of 1.5g. The output is
filtered using a n RC low pass filter with values R=1 K ad C=0.10.1μf. The output of the
sensor during typical seizure is shown on the graph. The output of MEMS is given to a 10 bit
analog to digital converter for digitizing the output.
(iii)Processor: The signals from sensors are processed using PIC18F4620 microcontroller. The
microcontroller requires a 10 bit ADC and a comparator circuit for processing the signals from
the sensor. PIC 18F 4620 includes built in ADC and comparator. The processor is clocked at
4MHz. The frequency of normal heart beat rate is 1.2 HZ. approximately. Or the time period of
the heart beat signal is 0.83 secs. The algorithm detects the sudden decrease in pulse width which
is one of the aura of epilepsy. As soon as the variations in the heart beat are detected, the
algorithm checks for the typical seizure waveform from the mems sensor. When these two
signals coincide, the software takes the decision as an epileptic seizure and generates control
signals.
(iv) Wireless transceiver: The device uses DATEx protocol for communication. The DATEx
Wireless Networking Protocol is a simple protocol designed for low data rate, short distance, and
low-cost networks. Fundamentally based on wireless personal area networks (WPANs), the
DATEx protocol provides an easy-to-use alternative for wireless communication. In particular, it
targets smaller applications that have relatively small network sizes, with few hops between
nodes, using Microchip’s MRF24J40 2.4 GHz transceiver for compliant networks. The DATEx
protocol is based on the MAC and PHY layers specification, and is tailored for simple network
development in the 2.4 GHz band. The protocol provides the features to find form and join a
network, as well as discovering nodes on the network and route to them. The card uses PCB
trace antenna. The device uses line of sight communication. The range is approx. 200ft. The
wireless transmitter and receiver hardware consist of a motherboard with PIC 18F4620
microcontroller. The motherboard consists of a daughter card with microchip MRF24J40 2.4
Ghz transceiver. The board is designed to work at 9V to 3.3 V DC. The DATEx protocol stack
can be installed on the board and the required application can be programmed into it. A peer to
peer network is formed using the transceivers.
One node is programmed as the network coordinator and the other as an end device. The
coordinator is set as the transmitter and the en device as the receiver. The long address is
assigned for the network and the short address to the nodes.
The device is designed in such a way that it searches for a network as soon as the module is
switched on. The coordinator assigns the address to the end devices and forms the network if
one is not detected.
The MRF24J40 is an compliant transceiver supporting DATEx, ZigBee™ and other proprietary
protocols. The MRF24J40 integrates wireless RF, PHY layer baseband and MAC layer
architectures that can be combined with a simple microprocessor to apply low data rate to a
multitude of applications The MRF24J40 device integrates a receiver, transmitter, VCO and PLL
into a single integrated circuit. It uses advanced radio architecture to minimize external part
count and power consumption. It mainly consists of TX/RX FIFOs, a CSMA-CA controller,
super frame Constructor, receive frame filter, security engine and digital signal processing
module. The MRF24J40 is fabricated by advanced 0.18 μm CMOS process and is offered in a
40-pin QFN 6x6 mm2 package.
The MRF24J40 consists of four major functional blocks:
1. An SPI interface that serves as a communication channel between the host controller and the
MRF24J40.
2. Control registers which are used to control and monitor the MRF24J40.
3. The MAC (Medium Access Control) module that implements compliant MAC logic.
4. The PHY (Physical Layer) driver that encodes and decodes the analog data. The device also
contains other support blocks, such as the on-chip voltage regulator, security module and system
control logic.
4. Design of software
The processing unit utilizes the logic implemented in the software for accurate detection of
seizures. The software checks the input signal from the PCI DIO 32 HS from the patient’s body
continuously and measures the pulse width of the signal. This width is converted into heart beat
rate by the
software. If there is any abnormalities in heart beat, it can be detected as a change in the pulse
width .As soon as the logic detects a change it triggers the vibrator and the system waits for the
response. The patient has to press a button on his wearable unit. If the patient is unable to do so
due to
occurrence of seizure, then response signal from MEM sensor which senses the muscular
convulsions is captured and analyzed. If there are signals of muscular convulsions the software
concludes that the patient has seizure and warning message is transmitted using the wireless
transmitter. The
seizure detection algorithm from the MEMS signals is to check only the sudden abnormality
occurring in the human body. This algorithm helps to avoid situations where heart beat rises due
to excessive physical work or due to tension etc. The algorithm uses the averaging technique to
determine
abnormalities accurately.
P=(P+N)/2
where p=previous heart beat rate
N=next heart beat rate.
For a person suffering from epilepsy, in the pre ictal stage the heart beat varies abruptly and
hence the value of P also changes. This change in the value of P is detected and the program is
made to wait for the signal from the second sensor which senses the muscular convulsions. If
muscular convulsions are detected from the second sensor, it triggers the transmitter on which
transmits a coded signal which is received by the receiver. The software section contains the
following major functional modules:
1. Heart beat rate calculations
2. Seizure detection from MEMS signal
3. Communication control
4. Overall supervision
5. Implementation
The system requires a heart beat sensor, muscular convulsion sensor, a transmitter, receiver,
mobile messaging device, alarm device and automatic vehicle control system. All the above said
parts are integrated together to a processor to form the device. The epilepsy prediction system
can be practically implemented by incorporating the following components:
a) Heart beat sensor: A pulse oxy meter is used as a heart beat sensor. The implementation of
PCI DIO 32 HS is by cascading several stages as shown in the figure 4. A high pass filter is
designed with lower cut off frequency of 15 Hz. .the high pass filter is cascaded with a low pass
filter designed to an upper cut off frequency of 0.5 Hz. The amplifier at the final stage raises the
voltage from mV level to the required voltage range. An amplification factor of 50 is given to it.
b) Convulsions sensor: An accelerometer is used as a convulsion sensor. Muscular convulsions
are detected using single axis mems IC MMA1260EG. The sensitivity of the sensor is set to
1.55g. The circuit is implemented as shown in the circuit diagram. The output of the sensor is
filtered out sing a low pass RC filter externally. The value of R is selected as 1K and C as
0.1μf.
c) Processing unit: The processing unit contains PIC 18F4620 microcontroller which is clocked
at 40 MHz.. PIC18F4620 have 64 Kbytes of Flash memory. The microcontroller has inbuilt 10
bit ADC which is used to digitize the output from MEMS module. It also includes a comparator
which is used to process the heart beat waveforms from the pulse oxy meter. The incoming
signal is processed using logics implemented in the software which runs the device. The
processing unit continuously checks for symptoms in the incoming signal. As soon as it detects
any abnormality, it triggers a warning
vibrator and the wireless transmitter.
d) Wireless Transmitter and receiver: Wireless transceiver consist of a board consisting of
MRF24J40 IC The transmitter transmits a coded signal which is decoded by a receiver to
generate control signals. The control signal activates an alarm device, mobile messaging device
and automatic vehicle control system appropriately. Apart from the above important blocks, a
buzzer circuit and a DC to DC convertor blocks are also implemented.
e) Enclosure design: The device is a wearable one (on the wrist). Hence the
enclosure is designed suiting to that purpose. The enclosure can be designed in the form of a
watch.
6. Software tools used:
1. MPLAB Integrated development Environment
2. Microchip C18 compiler
3. Keil Integrated development environment
7. Testing
(i) Testing of PCI DIO 32 HS: The PCI DIO 32 HS was tested by wounding the probe of the
device on the index finger of a person and the output were viewed on a DSO. The output is
shown in the graph given below. The PCI DIO 32 HS successfully detected the heart beat
waveform from the patient’s index finger. The out put frequency was 1.2 Hz . And the voltage
level was in the range of 100 to 120 mV.
(ii) Testing of MEMS sensor: The MEMS sensor is connected to the body of the patient using
straps. Typical epileptic seizure waveform is shown in the
figure below. This stage is not yet fully tested and testing is under way.
(iii)Testing of software: The inputs from the sensors were provided to the PIC controller in
which the software was programmed. Wave forms describing different conditions of the patient
were given as input and tested .
(iv)Testing of communication module The transceiver is directly connected to the
microcontroller in which the software was programmed. As soon as the software detected the
epileptic symptom, the transmitter was triggered. Using Zena network analyzer, the network was
detected at a frequency of 2.4G Hz. A peer to peer single node network was formed which
transmitted the message to the receiver node. The system designed here processes the heart beat
continuously and abnormalities are detected accurately. The device transmits the signal only
when seizures of epilepsy are detected. The performance of the device is not restricted by
movement of the patient. By using this device the patient can move freely without worries.
8. Problems encountered
We have encountered many problems as noted below:
1. Non availability of 3 axis accelerometer: We could not procure the 3 axis accelerometer and
hence testing is only performed with a single axis accelerometer. However, the system gives
better results only if a 3 axis accelerometer is used in for detecting muscle contractions.
2. Noise and temperature effect on the sensor outputs: Major problems were encountered due to
noise picked up by the sensors. Use of shielded cable and grounding solved the problems to a
satisfactory level. Heating effect of active components like op amps also created problems like
drifting and thermal noise. This was solved by operating the op amps at a lower voltage.
3. Problem with suitable wearable enclosure: A suitable wearable enclosure is not designed.
Compact PCB must be designed to fit all the components inside a wearable enclosure.
9. Advantages and benefits
The benefit of the project is that a lightweight, rugged, lowcost, wearable (on the wrist) device is
developed which helps a victim of epilepsy to do all sorts of activities like others do.
The device will be extremely cost effective since it uses simple sensors and technology for the
detection.
The sensors are small in size and can be firmly attached to the body.
Batteries can last long as the device consumes only little energy.
The device doesn’t restrict the movement of the patient.
The system is easily expandable paving the way to incorporate much more sophisticated
devices like ECG detector in the future.
Stand alone application.
10. Improvements
The system is easily expandable to incorporate GPS system and to capture and transmit various
patient parameters like ECG , body temperature etc.
11. Conclusion
A light weight, rugged, cost-effective wearable device is developed which helps millions of
victims of epilepsy around the globe. With the device in possession an epilepsy victim can move
around freely like normal people sans worries.
No comments:
Post a Comment