FPGA-Based Low-Power Architecture for Real-Time Arrhythmia Detection Using Fixed-Length Segmentation

Data processing and low power consumption are essential for monitoring heart signals in real time[5],[7].This project is a prototype of a real-time electrocardiogram (ECG) monitoring application for the Xilinx PYNQ-Z2 System-on-Chip (SoC) board[5],[10].The board's built-in 12-bit ADC converts signals from the heart[4],[6], which are picked up by using an AD8232 sensor, into digital values.High speed, high stability baseline wander removal[7],[9], electrical noise filtering and other heavy signal cleaning jobs are executed directly in the FPGA, which is a 25-tap FIR filter.The cleaned signal is then passed on to the ARM processor on board where the heartbeats are detected and the relevant points of P, Q, R, S and T waves are identified within a 5 second window in less than 2 milliseconds [1],[8].During the 89.7-second continuous live test, the system was able to accurately track the heartbeats at 75 bpm (800ms intervals) at 99.68% sensitivity and 99.74% positive predictivity [2],[3],[10].It consumes only 14.2% of logic resources, 3% of memory, and consumes approximately 160 mW of power, and it is considered as a working, lowpower hardware prototype [4],[7],[10].