FPGA Based Defibrillator Pulse Simulation Using ECG Signal Detection

Abstract

This work proposes a hardware–software co-design architecture for real-time electrocardiogram (ECG) signal transmission and classification using FPGA technology and embedded machine learning. Initially, the ECG signal is preprocessed in MATLAB, where the biomedical waveform is converted into hexadecimal two’s complement representation and stored as a memory initialization file. The formatted data is then imported into Block RAM (BRAM) using Xilinx Vivado IP and deployed on a Basys3 FPGA development board.A custom UART transmitter module is implemented to sequentially fetch ECG samples from BRAM and transmit them through serial communication at a baud rate of 9600. On the receiving end, a Raspberry Pi collects the incoming data via UART and reconstructs the 16-bit signed ECG signal. The recovered waveform is subsequently analyzed using a Random Forest–based machine learning model to determine whether the ECG pattern corresponds to a normal or abnormal condition.

The proposed design demonstrates efficient integration of FPGA-based digital hardware with embedded intelligence for biomedical signal processing. The architecture ensures reliable data transmission, reduced hardware complexity, and scalability for advanced signal analysis. This framework highlights the effectiveness of FPGA–embedded platforms for real-time physiological monitoring and intelligent healthcare diagnostics.