Design of Programmable Gaussian-Derived Wavelet Filter for Wearable Biomedical Sensor

Abstract

To provide multiple options for specific application in bio-signal processing, the programmable Gaussian-derived Gm-C wavelet filter has been proposed. To realize the programmable characteristic, the analog wavelet base with one numerator term is constructed by using hybrid artificial fish swarm algorithm. Also, the inverse follow-the-leader feedback Gm-C filter structure with a switch array is employed. By programming switches only, Gaussian and Marr wavelet transforms can be realized flexibly with all component parameters unchanged. The seventh-order programmable wavelet filter is designed as an example. Simulation results show that power consumption is only 141.68 pW at scale a=0.1, with dynamic range of 42.6 dB and figure-of-merit of 2.05×10-13. Due to the programmability, the proposed design method can implement two wavelet filters with very low circuit complexity.