A Memristor-Based Bipolar Cell Neuron Circuit for Motion Detection and Action Prediction

In this work, inspired by biological vision systems, a memristor-based bipolar cell neuron circuit (BCNC) is proposed, which can efficiently process incremental information to accurately encode changes in analog signals, improving the systems ability to perceive direction information of object movement. To further verify the feasibility, a BCNC based sensory nervous system constructed for detecting motion and predicting actions. Simulation results demonstrate that compared with traditional LIF neurons, BCNC has enhanced performance (14.2%) in accurately identifying motion and predicting actions in unmanned aerial vehicle (UAV) navigation tasks. This work provides a new idea for enriching the function of memristor based neuron circuit, and provides a new solution for the development of artificial neuron circuit with better biological similarity, and lays a foundation for the development of more intelligent artificial neuromorphic computing systems.