An Optoelectronic Memristor for Near‐Sensor Computing

ABSTRACT The optoelectronic memristor (MR) is a novel type of MR jointly affected by electrical and optical stimuli, offering diverse control methods and application scenarios. However, due to the lack of optoelectronic MR models and their equivalent circuit realizations, research on their hardware circuit application scenarios remains insufficient. To address this gap, we designed an optoelectronic MR model with an additional optically controlled terminal, incorporated an optical bias with electrical control to modulate the memductance of the model, and realized a floating equivalent circuit model for the optoelectronic MR. Our analysis clarified typical nonlinear memristive characteristics exhibited by the optoelectronic MR under different electrical and optical control signals. The validity and practicality of the optoelectronic MR model were thoroughly validated through a multifaceted approach that includes theoretical analysis, simulation analysis, and hardware experiments. In order to further demonstrate the potential applications of optoelectronic MR models, we innovatively utilized the recovery mode of memductance to identify input analog pulse sequence signals with different amplitudes as equivalent representations of binary “0” and “1” states combinations, and to extract the analog amplitude features of accumulated analog input signals, through volatile and nonvolatile optoelectronic MR models. Optical control provided additional regulation and enhancement to the processed results. We mark the first innovation in analog signal processing using the optoelectronic MR model for near‐sensor computing. The model significantly contributes to research on optoelectronic MR applications and their nonlinear characteristics, serving as a valuable reference for modeling and application research of other novel MR devices.