WO3−x:N/ZnO:N Broadband Optoelectronic Synapse for Object Detection in Edge Computing
Broadband optoelectronic memristors with high computational efficiency and low power consumption are pivotal for neuromorphic computing at the edge. This work presents a Ag/WO3−x:N/ZnO:N/ITO memristor exhibiting dual-modal synaptic plasticity. Electronic synaptic properties emulate biological plasticity, while photoresponse to multiple wavelengths, including simultaneous dual-wavelength stimulation, yields composite photocurrents. Leveraging these characteristics, single- and dual-wavelength artificial vision arrays simulate human visual perception. An artificial neural network integrated with a Field Programmable Gate Array (FPGA) forms a floating-point arithmetic system for object detection. The edge computing system achieves a 103-fold reduction in power consumption, addressing computational power limitations and enabling floating-point operations in embedded neuromorphic deployments. This work advances broadband optoelectronic synapses for efficient, low-power edge computing.