Accuracy Analysis And Control Of Multi-Axis Micro Assembly System Base On "PC+UMAC"

Micro-small parts have the characteristics of small geometry and high assembly accuracy,and it is difficult for the traditional manual assembly to fulfill the assembly requirements.With the development and application of computer vision technology,micro-assembly field has entered the stage of human-machine cooperation.In the assembly process,it is necessary to adjust the position of the parts in the operating space,more degrees of freedom are involved,and because of the combination of vision technology and micro-assembly,the vision error must be constrained in the range of error requirements.This paper takes the High-precision Alignment Detection System(HADS)as research platform and the main contents are as follows:(1)Analyzing the features of the HADS,discussing the pros and cons of the overall framework of the existing control system,combined with the specific assembly requirements,using the "PC + Independent Controller + UMAC" framework to complete the control system of HADS,completing the design of hardware system and electrical system.On the basis of hardware system,from the perspective of technical requirements of high precision alignment detector assembly process,analyzing the application layer and classification of the control system.In the executive layer of software,completing the system initialization operation and interlocking ware complete by programing the PLC.With the method of modularization,completing the software design of host computer control system.(2)Based on the principle of coaxial alignment detection,combining with the composition of micro-assembly subsystem and assembly flow,analyzing the error sources of vision module and proposing feasible control method.Among them,the effect of light parameters on visual error was analyzed based on the illumination model.A feasible optimization scheme was proposed to improve the image quality and to reduce the detection accuracy by adjusting the light source angle,brightness and camera exposure time.Based on camera imaging model,the existing calibration algorithm and pixel equivalent calibration method are analyzed and evaluated,and proposing a visual calibration error control method for calculating pixel equivalent calibration error by using appropriate high-precision calibration algorithm and capturing chessboard image which contains of a large number of corners.(3)The internal and external parameters of the camera are calibrated by Zhang’s plane calibration method using MATLAB.The deviation of the re-projection was sought out by using the method of re-projection.Using MATLAB to check the checkerboard calibration points and calibrate the pixel equivalent value.Using the method of re-projection to calculate the error mean.The effects of light angle,light intensity and exposure time on the imaging effect were analyzed by orthogonal experiment,and the optimal parameters of primary reflection and secondary reflection were obtained by combining high precision array hole and laser confocal detection.Using the method of inverse assembly to estimate the approximate error of the system.Based on the above research,the hardware construction and software design of the high precision alignment detector are completed,and the error of the visual part of the system is analyzed and the related control method is put forward.