Sensorless Control of Surface-Mounted Permanent Magnet Synchronous Motor Using Adaptive Robust UKF

With the increasing demand for sensorless control of permanent magnet synchronous motor (PMSM), unscented Kalman filter (UKF) is widely used in sensorless control of PMSM, as it overcomes the shortcoming of neglecting higher-order terms in the linearization process, and improves the calculation accuracy of nonlinear distribution statistics. However, UKF is susceptible to system noise and measurement gross errors, which limits the application of UKF. In this paper, an adaptive robust UKF algorithm is proposed and studied, which combines the advantages of robust Kalman filtering and adaptive Kalman filtering. Based on the principle of the equivalent weight of robust M-estimation, the equivalent covariance matrix is obtained, and the observation gross error of the system is reduced. Meanwhile, Sage-Husa estimation algorithm is adopted to adjust the system and process noise adaptively. Finally, the adaptive robust UKF is applied to the sensorless control of surface-mounted PMSM (SPMSM), the simulation model of the proposed adaptive robust UKF and traditional UKF method are established and solved in MATLAB/Simulink respectively. To verify the effectiveness of the adaptive robust UKF algorithm, an experimental test platform is built and the experimental tests are carried out. Compared with the traditional UKF method, the simulation and experiment results show that adaptive robust UKF algorithm has smaller estimation error of rotor speed and position, higher robustness in the case of speed jumping and load disturbance.