Multi‐Innovation Kalman Sliding Mode‐Based Model‐Free Adaptive Control for PMSMs

ABSTRACT Permanent magnet synchronous motors (PMSMs) are widely adopted in high‐performance applications such as electric vehicles, industrial robotics, and precision machine tools, where stringent speed regulation is required despite uncertainties in operating conditions and system parameters. To overcome these limitations, this article presents a multi‐innovation Kalman sliding mode‐based model‐free adaptive control approach. First, a full‐form dynamic linearization‐based MFAC architecture eliminates reliance on precise physical parameters while preserving system dynamics through pseudo‐gradient estimation. Second, a Kalman filter‐embedded adaptive sliding surface is theoretically derived to suppress chattering phenomena while maintaining strong robustness against uncertainties; the equivalence between Kalman filtering and optimal disturbance attenuation is rigorously established. Third, multi‐innovation identification theory is incorporated to optimize the sliding surface adaptation mechanism, wherein historical data vectors enhance estimation accuracy and convergence speed through recursive innovation exploitation. Fourth, a state‐dependent adaptive exponential reaching law is designed to accelerate sliding mode transition while eliminating reaching phase chattering. Comprehensive simulations and experimental validations on a PMSM test bench demonstrate that the proposed method achieves superior tracking accuracy, faster dynamic response, and enhanced robustness.