A High Speed and High Accurate Floating-Point EEG Signal Bandpass Filter Based on FPGA
Zhaoping Zeng, Liangtao Yang, LI Guangbin, Zhilin Zhang, Yi Zhang, Mingqiang Huang, Jinglong Wu, Yutang Li, Chunlin Li · 2023
The filtering of different frequency bands plays a crucial role in the EEG signals processing. Although existing FPGA based EEG signal filtering have good processing speed, the floating-point signal processing and parallel computing are still challenging. Herein, we propose a one-dimensional digital bandpass filtering method based on unfolding data, which not only possesses low resource consumption but also has outstanding computational reconfigurability. The efficient EEG signal filtering algorithm is first implemented in MATLAB, and then an FPGA-based bandpass filter is designed, so that floating-point EEG signals can be processed in miniature hardware. Our design achieves a low precision loss of 0.02% -2% with a 150 times acceleration compared to the conventional platform.