摘要

With the advantage of low cost and small size, a hall sensor has been exploited to detect motor position, and this strategy has attracted much attention. A challenge is that measurement model parameters vary with position and vary from motor to motor because of the machining errors of permanent magnets. While, most existing methods use other auxiliary displacement sensors to calibrate model parameters off-line, which require a lot of preliminary work and lose adaptation when they are applied in industry. In this paper, an on-line simultaneous computation of model parameters and position method is proposed to measure position for a permanent magnet synchronous linear motor. The measurement model with Fourier series is employed to approximate the mapping relationship between magnetic field strength and mover position. Based on the continuous variation characteristics of model parameters and position, a simultaneous computational algorithm is proposed to real-timely compute model parameters and position without other sensors. In the experiments of one-way and reciprocating motion, magnetic field information is detected by a self-designed hall sensor array. Experimental results indicate that the standard deviation of measurement error is less than 10 mu m with the comparison of a reference optical encoder.