Resistive switching memory based on polyvinyl alcohol-graphene oxide hybrid material for the visual perception nervous system

Abstract

Resistive random-access memory (RRAM) is a new memory technology that can not only realize high-density storage, but also can simulate the neural synapse for use in artificial intelligence applications. In this study, we propose an RRAM device that shows competitive resistive memory characteristics and can similarly be used as a synapse in simulation of the human visual perception nervous system. First, we demonstrate that the polyvinyl alcohol-graphene oxide (PVA@GO) hybrid material-based RRAM device offers competitive resistive memory characteristics, including long retention capability, high durability, repeatability, and mechanical flexibility. Second, we integrate the RRAM (as the artificial synapse) with a light-sensitive electronic component (a photoreceptor cell) to construct an artificial visual perception system, and realize effective emulation of light perception and conversion of light signals into synaptic signals. Under light irradiation at 532 nm, a range of versatile synaptic functions, including short-term plasticity (STP), long-term plasticity (LTP), and paired pulse facilitation (PPF), was imitated. This work provides valuable insight into the development path for next-generation high-density data storage technology, and also offers a new way to imitate the human visual neural network for multi-functional humanoid robots.

Keywords

• RRAM
• Graphene oxide
• Artificial synapse
• Visual perception nervous system