Sensorless drive of permanent magnet linear motors using modified Kalman filter

This paper presents a novel method to eliminate the need for a position sensor in permanent magnet linear motor (PMLM) drives, by using modified Kalman filter estimation algorithm. Due to end effects, the PMLM must operate at high speed dynamics, and its lowest speed can go down to zero velocity. This problem is overcome by using sinusoidal open loop current control for speeds below 0.1 m/s. The Kalman filter is engaged when the speed goes above 0.1 m/s. In the investigation, the control and estimation algorithm, together with the PMLM are simulated in MATLAB and SIMULINK. After that, the drive system is implemented in hardware with a DSP controller board. Results show that the Kalman filter is capable of estimating system states accurately and reliability without mechanical sensors. Together with the low-speed open-loop control, the PMLM can be operated under trajectory tracking mode without any position sensors.