Precise-robust position measurement of a linear motor based on the extended speed-up robust features algorithm and sawtooth stripe pattern.

Abstract

To improve the accuracy and robustness of linear motor mover position detection, a linear motor displacement measurement method is proposed based on the extended speed-up robust features algorithm and sawtooth stripe image. First, a sawtooth stripe image is constructed as the target image. To optimize the target image with strong robustness, the spatial frequency and image standard deviation are introduced as the image quality evaluation indices. Second, a line scan camera fixed on the linear motor is used to capture the target image signals in real time. The sequential sawtooth stripe signals are preprocessed by filling sampling to improve the matching rate of feature points. To satisfy the real-time requirement of mover position detection, the singular value decomposition is used to reduce the dimension of the preprocessed image. Subsequently, an improved speed-up robust features algorithm is used to achieve sub-pixel displacement measurement. Finally, the actual displacement of the mover can be calculated by the calibration coefficient of the measurement system. Simulation and comparative experiments show that the proposed sawtooth stripe image has better robustness, in contrast with the fence image and aperiodic sinusoidal image in references. It is also demonstrated that the proposed method has higher accuracy and anti-interference performance than other methods under different conditions.