PMSM Control Strategy Based on SVDGHCKF

In the study of sensorless control strategies for PMSM, this paper proposes a generalized higher-order cubature Kal-man filtering algorithm based on singular value decomposition, to address the limitations of the traditional CKF observer's low observation accuracy and filter divergence. The algorithm tries to improve the PMSM's control performance. First, this work establishes an accurate mathematical model of the PMSM as a research foundation. The classic CKF algorithm's observation accuracy is then greatly improved by introducing the fifth-order spherical radial volume criterion, which raises the observation ac-curacy to the fifth-order level. Meanwhile, to address the filter dispersion issue, this research uses the SVD decomposition to replace the Cholesky decomposition in the classic CKF. Finally, the research indicates that the control strategy can achieve high-precision control of the speed and angle of the permanent magnet synchronous motor, which effectively improves the control performance of the PMSM.