Extended Kalman Filter as the Prediction Model in Sensorless Predictive Control of Induction Motor

The accuracy and robustness of the prediction model are always critical issues in model predictive control (MPC). This is more serious in sensorless applications because there are more uncertain parameters in the control system. Extended Kalman filter (EKF) is known as one of the self-correction methods. It has been widely used in sensorless applications as the observer with the aim of speed estimation. In this research, a new prediction model based on EKF is proposed and studied. This study aims to investigate the effectiveness of the EKF-based prediction model in the presence of parameter mismatch in the sensorless application of the predictive method. The simulation results verify the validity of the proposed method.