Position Calibration techniques Comparison for Sensorless Controlled PMSM

Two calibration techniques (stochastic variance estimation and angle correction) for position estimation in sensorless controlled permanent magnet synchronous motor (PMSM) are compared. Four sensorless estimations were employed in the experiment. These include flux-linkage observer (FLO) based on direct calculation without correction, extended Kalman filter (EKF) and linear Kalman filter (LKF) observation based on variance convergence and single dimension Luenberger (SDL) estimation based on angle correction. It was concluded that variance convergence is still best choice for state estimation.