Battery-Free Neuromorphic Computing Based on PMMA/SiO2 Nanoparticle Memristor and Near-Field CommunicationClick to copy article link
The human intelligent perceptional system employs two fundamental modules─sensing and computing─to engage with the external environment. The von Neumann architectures, however, inherently face the memory-wall and power-wall challenges. In response to these issues, neuromorphic computing has emerged as a potential paradigm, minimizing data migration via in-memory computing frameworks and eliminating analog-to-digital conversion through near-sensor processing. Despite this, conventional methods of power supply and data transmission constrain the application of this nanoscale neuromorphic architecture in edge computing. This article presents a battery-free neuromorphic computing approach that integrates memristor-based neuromorphic computing with near-field communication (NFC) transponder chip-assisted power supply and wireless communication. The performance of this system is evaluated based on its ability to classify distinct optical characters. The battery-free power supply ensures energy consumption only during the execution of the classification task, while neuromorphic computing reduces the volume of data to be transferred. Our demonstration offers a promising avenue for the application of neuromorphic computing, particularly in the field of edge computing.