Memristor-based neural network PID controller for buck converter

This paper applied spintronic memristors on the neural network PID controller to control the nonlinear and time-varying buck converter. The proposed method used spintronic memristor as electronic synapses to store the updated value, which can greatly simplify the implementation of hardware circuit and the weights updating algorithm. Meanwhile three-layers memristor-based BP neural network was designed and the algorithm of the PID control under this mechanism was also analyzed. Furthermore, computer simulation of buck converter working in continuous mode with the proposed controller presented superior performance in the aspects of follow-up, immunity and robustness, as well as fast response and small overshoot.