A bio-inspired ultra-low-power spike encoding circuit for speech edge detection

Speech edge detection can be used to adaptively control the performance and improve the energy efficiency of smart audio sensors. In this paper, we propose a spike encoding circuit inspired by mammalian auditory system for real-time and low-power speech edge detection. The circuit can directly encode the signal's envelope information by asynchronous spikes' temporal density without additional envelope extraction. Furthermore, the spike encoding automatically adapts its encoding resolution to the amplitude of the input signal, which improves encoding resolution for small signal without directly increasing power consumption. Fabricated in 0.5-μm CMOS process, the spike encoding circuit consumes 0.3-μW power, and the experimental results are presented.