Self-Powered Synaptic Transistor for Artificial Perception

Abstract

Artificial perception system is expected to perceive and integrate a large amount of sensory data to dynamically train inspired neural networks. It is expected to become a promising candidate for the realization of new imitating neural networks. But all reported artificial perception usually requires additional voltage to operate the tactile sensor, and most of the research on tactile sensing lacks learning and memory functions. Here, we propose a self-powered artificial perception system, which is combined with triboelectric nanogenerators (TENG) and synaptic transistors to realize a unique self-powered artificial synapse system. It successfully demonstrates the typical characteristics of sensory memory, such as excitatory postsynaptic current and paired impulse promotion. The impulse signal response under different pressure senses is created. Finally, Ebbinghaus memory is also simulated by the system. The artificial nervous system is expected to become a promising technology for neural perception and artificial skin systems in the future.