Abstract: Aiming at developing an effective and reliable sensorless control algorithm for permanent magnet synchronous motors， for the equivalent circuit model considering iron loss resistance， the differential expression of the torque current state was derived under the d-q rotating axis system and the α-β stationary axis system. On this basis， a sliding mode observer was designed to estimate the speed and rotor position signals. In order to verify the observation effect， the sensorless algorithm was simulated and compared under the two axis systems. A fusion observation algorithm based on the extended Kalman filtering was proposed according to the characteristics of the low accuracy of the rotor position signal under the d-q axis system and the high frequency chattering of the signals under the α-β axis system. Simulation results showed that the fusion observation algorithm has less error and good dynamic characteristics. Besides， the effectiveness of the fusion observation algorithm was verified by experiments.

Key words: permanent magnet synchronous motor(PMSM); iron loss resistance; sensorless control; extended Kalman filter; fusion observation algorithm