BME 458 – Group M6
Ryan Garrone
Gu Eon Kang
Eunjoo Hwang
Group M6: InRhythm
Final Design Project
Introduction
Arrhythmia is defined as an irregularity in the force or rhythm of the heartbeat. Arrhythmias
contribute to many health problems and complications, such as sudden arrhythmia death syndrome
(SADS), which causes over 300,000 deaths per year in USA alone. The objective of our project is to
build an automatic IV drip system that continuously monitors a patient’s heart rate, logs any issues that
are detected, and maintains the IV fluid close to body temperature. Since SADS related deaths can be
reduced by 25% with early diagnosis, our system could serve as an inexpensive monitoring tool to notify
doctors of any arrhythmias before major complications, such as cardiac arrest, occur. In addition, current
diagnostic methods are mostly expensive and requiring hospital stays, so our IV drip system enables
patients to continue their daily life normally, without having to be stuck in a hospital for diagnostic
purposes. Our IV drip system responds to a patient’s heart rate by looking at the distances between
peaks. If our system detects either Bradycardia or Tachycardia, the IV drip will automatically begin to
flow until the patients heart rate goes back to normal. Our IV fluid will be continuously monitored with a
thermistor and its temperature will be controlled using a peltier plate. We hope that our overall system
will provide an inexpensive yet accurate method for doctors to help diagnose arrhythmias and prevent any
complications by providing feedback to the doctor regarding the patient’s heart condition.
Experimental Setup
Our setup was composed of both hardware and software components. For all of our software, we
used LabVIEW to do the data processing for our system. Our LabVIEW VI consists of two sub-VIs, one
which acquires heart rate and the other which acquires temperature. Both sub-VIs take voltage
measurements from the DAQ. First, the heart rate is obtained via ECG signals. We filter out 60Hz noise
and eliminate the DC offset further to give us a cleaner signal. We measured the period between one peak
and next peak to calculate heart beats per minute and used an average over time to distinguish between a
sudden change in heart rate change due to arrhythmia and normal physical activity. For our temperature
measurement, the voltage was converted into temperature via a linear calibration function. Both the
temperature and the heart rate were used to activate the IV drip and peltier plate by comparing threshold
and current values with a logic function. The digital waveform output function of LabVIEW was used for
switching the two on/off by sending 5V square waveforms to the outputs. Overall, the software was used
to monitor and process signals and control our hardware.
For the hardware side of our system, there were a few major subsystems. First was our
temperature control subsystem, which was composed of a thermistor, a peltier plate, a heat exchange
system, and a MOSFET switching circuit. There was a heat sink for both the hot and cold side, and a fan
aided in the heat exchange between the fluid and the peltier plate. The temperature subsystem was
directly mounted to our rigged up IV “bag,” which was actually just a plastic container with a heat sink
glued into the bottom due the our budget limitations. The next subsystem was our solenoid switching
valve, which controlled the IV fluid flow. It consisted of our solenoid valve and its MOSFET switching
circuit. Both the valve and the temperature control subunits were controlled via LabVIEW. Finally, we
had our ECG subsystem. This was composed of our three electrodes (ground, left arm, and right arm) and
our filtering circuitry. Our filtering circuitry consisted of a pre-amp to provide gain and remove noise,
two cascaded second order Sallen-Key filters which served as our band pass filter (0.5 Hz to 10 Hz), and
a high pass filter to provide more gain and reduce the DC drift further. This combined set up allowed our
system to function as a unit that had both monitoring and controlling functionality.
Results
Our system functioned very well once all of our hardware issues were sorted out. In order to
meet our acceptance criteria, we wanted to be able to get a high quality ECG signal, accurately monitor
and control fluid temperature, control the IV drip based on the heart rate, and send notifications to doctors
based on patient problems. In the end we were able to meet all of the above criteria with our system.
For our heart rate monitoring, we were able to produce the results below, which show accurate
ECG signals and the correct calculation of both a resting and an elevated heart rate, with a corresponding
warning bar. Triggering
the solenoid valve was
simply a logical
comparison with a
threshold and the current
heart rate, and so once
the heart rate exceeded
our set threshold of 100,
our solenoid opened,
allowing fluid to flow
down the IV drip.
In order to monitor our fluid’s temperature, we used a thermistor, which changed its resistance
linearly with temperature. This allowed us to measure the changing voltage across it in a voltage divider
set up. We see from our calibration graph below that our linear model very closely approximated the
actual temperature by using the thermistor.
Regarding the
controlling of the temperature of
the fluid, we were able to
maintain the fluid at roughly 80
degrees Fahrenheit. This was
less than our desired body
temperature, however we were
happy with our results because a
more robust system would have
had much better thermal contact
and insulation, which would
allow us to control the
temperature of the fluid much
more. For our small budget, this
was acceptable as a proof of
concept since we could control the temperature to up to this threshold.
One area we had initial issues with was our doctor notification subsystem. Due to network
firewall problems at the university we were unable to send alerts with our VI on campus. We were
however able to validate our concept by running the VI at home, which sent both a text and e-mail to our
“doctor” notifying them that there was a detected problem with the patient’s heart. We also incorporated
a web monitoring system which would allow anyone on the same network to view the VI in real time.
We thought that this would allow the doctor to check up on patient’s hearts from time to time by seeing
real time wave-forms.
Our system as a whole was very reliable. Once our filters removed the drifting DC baseline, the
signal was very stable and we were able to get consistent measurements of heart rate. The temperature set
up was also very robust and we had no issues with it. One area of concern in the temperature system is
that the measured voltage is dependent upon the applied voltage across the voltage divider, so the voltage
needs to be steady and constant in order to get an accurate reading. Overall though, we consider our
system to be very stable and reliable for a low budget prototype.
Discussion and Conclusion
From our results we were able to conclude that the ECG signal, once filtered properly, is a very
accurate way to measure heart rate and associated arrhythmias. By measuring the different distances
between peaks and wave parts, it is fairly straight forward to diagnose different heart conditions.
In our experiment there were a few sources of error, such as EMG signals picked up by the ECG
and a slight amount of 60 Hz noise. We were able to get around the EMG signals by keeping the patient
stationary, and the 60 Hz noise was much smaller than our measured signal, so we were able to ignore it.
More accurate calculations of the average heart rate would have been helpful, since arrhythmia conditions
are usually diagnosed by looking at the heart rate patterns over hours. Another source of error was our
hand-crafted IV “bag.” It lacked insulation and good thermal contact, so we weren’t able to keep it at
body temperature and had to settle for a temperature closer to 80 degrees Fahrenheit.
Since our system was a very low budget prototype, it obviously had areas for improvement. We
would like to first see the entire system become portable and lose its wired connections. This would
make it more practical for a patient to wear around on a regular basis. The next main area for
improvement would be in our IV “bag” and its temperature system. The peltier plate can both heat up
and cool down the unit, so it would be nice to incorporate both so that the fluid could be maintained at
any set temperature. Also, the IV “bag” would need to be more insulated to provide better temperature
stability and the thermal contact with the peltier plate would need to be increased with thermal paste and
better overall contact between the heat exchange units.
Application
The additional functionality that we would add to our system would be increased
portability by running all the processing on a microcontroller. This would actually allow people
to use this device 24/7 and not have to be tied to a bulky computer. More advanced computation
of heart conditions would also be beneficial. Simple analysis of distances between certain peaks
can be used to diagnose a plethora of conditions, so by adding in this capability, our system
would become much more versatile and useful for patients who had a wide variety of heart
issues. This would be beneficial to health care providers and patients by allowing them to have
an extremely low cost and portable heart monitoring system to replace the current systems,
which are multiple thousands of dollars and do not offer diagnosis, just ECG recording.