Session
Mechatronics, System Engineering and Robotics
Description
A surge in the yearning for portable gadgets related to healthcare has ignited a coupling of microchip controllers with models from the realm of machine learning, aimed at the task of immediate health observation. It's quite paramount, specifically in the context of identifying irregularities in heart rate, like tachycardia—where the heart races unusually fast—and bradycardia, an instance of a sluggish heartbeat. The neat thing is that this study takes a stab at contrasting the effectiveness among three popular microcontroller types—the ESP32, the Raspberry Pi, and the Arduino Nano 33 BLE Sense—in terms of real-time tracking of heartbeats and spotting any oddities (using, of course, a model developed with some machine learning elements). At the heart of this model lies the fundamental structure known as Long Short-Term Memory, or LSTM. This model has been finely tuned to identify those pesky heart rate inconsistencies, after which it finds its home, so to speak, on each of the microcontroller types where it's applied for immediate reasoning. Microcontrollers are assessed on some pivotal metrics, alright? Think inference time or latency; toss in power use, how precise they are, the simplicity of melding them with heart rate sensors, and all that costing - that's good to keep in mind. What we see from our research is that each of these little guys has its perks, certainly - but there's always that trade-off lurking about, with processing mightiness, energy economy, and system expenses. The findings from this mash-up analysis are worth their weight for picking microcontrollers in the healthcare arena, especially for those wearables that can't stop monitoring heart rates. And then, this exploration's deductions are slipping into making more efficient, precise, and costfriendly healthcare system monitoring stuff.
Keywords:
Heart rate monitoringmicrocontroller comparisonmachine learning in healthcare.
Proceedings Editor
Edmond Hajrizi
ISBN
978-9951-982-15-3
Location
UBT Kampus, Lipjan
Start Date
25-10-2024 9:00 AM
End Date
27-10-2024 6:00 PM
DOI
10.33107/ubt-ic.2024.205
Recommended Citation
Jetullahu, Arxhend; Mulaku, Luan; and Demolli, Saranda, "Comparative Evaluation of Microcontrollers for Real-Time Heart Rate Monitoring and Tachycardia/Bradycardia Detection" (2024). UBT International Conference. 4.
https://knowledgecenter.ubt-uni.net/conference/2024UBTIC/MSER/4
Included in
Comparative Evaluation of Microcontrollers for Real-Time Heart Rate Monitoring and Tachycardia/Bradycardia Detection
UBT Kampus, Lipjan
A surge in the yearning for portable gadgets related to healthcare has ignited a coupling of microchip controllers with models from the realm of machine learning, aimed at the task of immediate health observation. It's quite paramount, specifically in the context of identifying irregularities in heart rate, like tachycardia—where the heart races unusually fast—and bradycardia, an instance of a sluggish heartbeat. The neat thing is that this study takes a stab at contrasting the effectiveness among three popular microcontroller types—the ESP32, the Raspberry Pi, and the Arduino Nano 33 BLE Sense—in terms of real-time tracking of heartbeats and spotting any oddities (using, of course, a model developed with some machine learning elements). At the heart of this model lies the fundamental structure known as Long Short-Term Memory, or LSTM. This model has been finely tuned to identify those pesky heart rate inconsistencies, after which it finds its home, so to speak, on each of the microcontroller types where it's applied for immediate reasoning. Microcontrollers are assessed on some pivotal metrics, alright? Think inference time or latency; toss in power use, how precise they are, the simplicity of melding them with heart rate sensors, and all that costing - that's good to keep in mind. What we see from our research is that each of these little guys has its perks, certainly - but there's always that trade-off lurking about, with processing mightiness, energy economy, and system expenses. The findings from this mash-up analysis are worth their weight for picking microcontrollers in the healthcare arena, especially for those wearables that can't stop monitoring heart rates. And then, this exploration's deductions are slipping into making more efficient, precise, and costfriendly healthcare system monitoring stuff.