Neuromorphic Processing of Pressure Signal Using Integrated Sensor-Synaptic Device Capable of Selective and Reversible Short- and Long-Term Plasticity Operation

摘要

To mimic the tactile sensing properties of the human skin, signals from tactile sensors need to be processed in an efficient manner. The integration of the tactile sensor with a neuromorphic device can potentially address this issue, as the neuromorphic device has both signal processing and memory capability through which parallel and efficient processing of information is possible. In this article, an intelligent haptic perception device (IHPD) is presented that combines pressure sensing with an organic electrochemical transistor-based synaptic device into a simple device architecture. More importantly, the IHPD is capable of rapid and reversible switching between short-term plasticity (STP) and long-term plasticity (LTP) operation through which accelerated learning, processing of new information, and distinctive operation of STP and LTP are possible. Various types of pressure information such as magnitude, rate, and duration were processed utilizing STP by which error-tolerant perception was demonstrated. Meanwhile, memorization and learning of pressure through a stepwise change in a conductive state was demonstrated using LTP. These demonstrations present unique approaches to process and learn tactile information, which can potentially be utilized in various electronic skin applications in the future.