A Multimodal Artificial Sensory Neuron Realized by Bio-Voltage-Level Threshold Switching Memristor

Abstract

The rapid advancement of brain-inspired neuromorphic computing is now driving the development of efficient artificial sensory systems as a key research frontier. In this study, we propose a threshold switching memristor (TSM) with a structure of Pt/Ag/Al2O3/HfO2/Ag NIs/Pt with a bio-voltage-level threshold voltage, and design a highly integrated and low-power multimodal artificial sensory neuron circuit based on the TSM. The artificial sensory neuron circuit is capable of real-time sensing of pressure and temperature information and realizes for the first time the encoding of pressure and temperature changes into output spike amplitude changes. The experimental results demonstrate that we have successfully developed an artificial sensory neuron capable of multimodal perception and spike amplitude encoding, providing new insights for advancements in bionic robotics, intelligent healthcare, and edge computing.