Design and verification of neural network sliding mode controller based on FPGA

This paper presents a design and verification method for a neural network sliding mode controller based on Field Programmable Gate Array(FPGA). This method uses FPGA as the main method of designing neural network sliding mode controller. Design experiments to verify the feasibility of the analysis algorithm for the core control variable control law and PWM signal in the doubly salient electromagnetic generator system (DSEG system). By intercepting the corresponding experimental data of the state variables in the finite element simulation of the motor as input, the corresponding control law and PWM signal are output after calculation by the neural network sliding mode controller. At the same time, the image processing capability of LabVIEW is used to compare and analyze the control law obtained by finite element simulation with the PWM signal and the signal output by FPGA. The experimental results show that the FPGA core controller can follow the control law well whether the system is in steady state or dynamic state. Corresponding PWM signal change trend is also consistent, which can achieve a better restoration of the simulation control state.