Current control of steering feel feedback motor based on adaptive prediction correction and parameter robust identification method

This paper proposes a robust current control strategy for permanent magnet synchronous motors in steer-by-wire steering feel feedback systems under sampling errors and parameter variations. A residual-driven adaptive current prediction correction method is developed to dynamically fuse measured and predicted currents, thereby reducing the influence of sampling distortion. Meanwhile, a Polynomial Chaos Expansion-enhanced Adaptive Extended Kalman Filter is introduced to identify stator resistance and equivalent inductance online while characterising parameter uncertainty. Simulation results show that, under a 2 A sampling error, the proposed method reduces the average current oscillation amplitude by 56.09% and 56.29% compared with PI and deadbeat predictive control, respectively. Under parameter step variations, the torque tracking error index is reduced by 43.45% and 61.02%, demonstrating improved current robustness and steering feel torque fidelity.