Hardware Accelerator Design of Personal Identification with ECG Signal Based on Artificial Neural Network

A hardware accelerator of personal identification with ECG signal based on artificial neural network (ANN) is proposed. ECG signal is used as the training data for machine learning to implement personal identification. The accuracy rate of identification is as high as 98%. Using NIOS to transfer data to SDRAM on FPGA, and design an ANN acceleration circuit to implement identity recognition function. Compared with the CPU for the same operation, the proposed accelerator has the same accuracy and more than 50 times speed improvement.