Research On Position Detection Method Of Linear Motor Mover Based On Joint Spatial Phase And LUMLS

Permanent magnet synchronous linear motors(PMSLMs)have no intermediate mechanical transmission link,simple structure,fast acceleration,small steady-state error,good thrust quality and other obvious merits.They are often used in laser engraving machine,computer numerical control machine tool and other high-precision industrial fields.The real-time and accuracy of mover position detection is the key step of the high-precision control of PMSLM servo control system.The robustness of the servo system will be affected by the anti-interference of position detection.Traditional methods for detecting the mover position of PMSLM mostly use position sensors such as magnetic grid and grating,which have the disadvantages of easy demagnetization,high cost and poor anti-interference ability.The merits of digital image processing measurement method based on machine vision are non-contact,low installation requirements,high accuracy and large stroke measurement.It can be considered to be applied to one-dimensional rigid body translation measurement of linear motor.Therefore,the image matching algorithm is applied to mover positioning,which further meets the high accuracy and robustness,and avoids the shortage of position sensor.The current research work focuses on the implementation of image measurement methods.The existing algorithms have some disadvantages,such as low accuracy,high time complexity,small measurement range.Aiming to increase the accuracy,anti-interference and rapidity of PMSLM mover position detection,a sub-pixel measurement algorithm based on joint spatial phase(JSP)is studied by using line scan camera to obtain one-dimensional fringe image sequence in motion.To further improve the speed and accuracy of the matching algorithm,and make the gray level of a single fringe period of the target image have enough randomness,a local upsampling moving least squares(LUMLS)algorithm is proposed.At the same time,the aperiodic sinusoidal fringe signals are constructed as the processing object.Theoretical analysis,simulation comparison and experimental results all prove the proposed method.The specific contents of this research are shown as follows:1.The motion characteristics of the PMSLM are analyzed,the gray contrast of different aperiodic fringe images is studied,the spectral width of power spectrum is introduced to evaluate the image quality.The low-cost line scan camera is used to establish the model of the mover position measurement system.The key to realize the accurate positioning of linear motor lies in the measurement accuracy of displacement between adjacent image signals.2.Aiming to detect the mover position of linear motor more quickly and accurately,a sub-pixel measurement method based on joint spatial phase is studied.Firstly,one-dimensional image signals are taken as the detection object,and the improved spatial absolute difference sum algorithm is used for coarse registration.The overlapped regions are extracted according to the registration results to form two new images.Then,the spectrum leakage in Fourier transform is suppressed by the Hamming window function.The cross power spectrum is obtained using phase correlation method.The sub-pixel offset is obtained by the phase angle function without unwrapping and the random sample consistency method finally.The simulation results show that it can detect the mover position faster and more accurately,and the measurement range is larger and time-saving.3.Considering the requirements of high-speed measurement and anti-interference,a high-precision robust measurement method based on one-dimensional aperiodic sinusoidal fringe signal and local up sampling moving least square method is proposed.The local up sampling cross-correlation curves of adjacent image signals are calculated by matrix multiplication discrete Fourier transform.To improve the precision of measuring further,the moving least square method based on Gaussian weight function is used to extract more accurate sub-pixel displacement.Theoretical and simulation results present that this method is better than other algorithms in measurement accuracy,computational efficiency,robustness and measurement range.4.The experimental platform of PMSLM mover position measurement is established,and the correctness and feasibility of the method proposed in this paper are verified.The gray level uniformity of the line-scanning camera is corrected,and the reliability of the calculation is guaranteed.The tracking performance of this method is verified compared with that of grating.To further verify the superiority of the optimized aperiodic sinusoidal image constructed in this paper,three different fringe images are used for displacement detection.The experimental results present that the LUMLS method has better measurement performance,stronger stability and anti-interference,and can be used to measure the motor position in the actual environment.