60 nm Pixel-size pressure piezo-memory system as ultrahigh-resolution neuromorphic tactile sensor for in-chip computing

Abstract

Neuromorphic tactile sensors provide considerable feasibility to simulate essential capabilities of the human tactile system even beyond the bio-perceptible range. Nevertheless, the complicated circuitry of artificial tactile sensors based on electrical-skin, memory, and processing components show serial barriers in terms of system resolution, integration, and power consumption. Here, a high-resolution pressure piezo-memory system (HPPMS) is reported with 60 nm pixel-size as a competent neuromorphic tactile sensor, that presents non-volatile force resistive conversion and force-tunable synaptic functions. The piezo-memory pixel arrays realize nanoscale force-image sensing and memory actions as same as neuromorphic tactile preprocessing by enhanced processing efficiencies and recognition rates. The proof-of-concept system affords the possibility to achieve ultra-small force distributions, simplify tactile sensor circuits, and develop advanced applications in in-sensor computing and artificial intelligence.