Research And Implementation Of Shaft-Hole Assembly Based On Baxter Robot

Regardless of the matching accuracy,this paper is aimed at a generalized shaft-hole assembly,that is,the insertion work.In the traditional sense,the shaft-hole assembly is mostly carried out in the form of online teaching or offline programming,and often does not have good robustness and flexibility.With the development of industrial automation,the demand for machine vision is getting higher and higher,so the development of visually guided shaft-hole assembly technology has important production and scientific significance.For the research and implementation of shaft-hole assembly under visual guidance,the main work of the thesis is arranged as follows:First of all,the seven-degree-of-freedom Baxter robot is built in the V-REP simulation platform.The simulation process is designed and the simulation program is written.Therefore,the shaft-hole assembly simulation using one arm of Baxter robot is completed.In order to improve the assembly space,the shaft-hole assembly simulation using both two arms of Baxter robot is designed and carried out.Secondly,the pose control function is designed in order to realize the point-to-point operation of Baxter robot.According to this function,the shaft-hole assembly experiments are carried out using one arm and two arms of Baxter robot.The feasibility and accuracy of the pose control function are verified according to the experimental results.Thirdly,the shaft-hole assembly physical experiment based on two-dimensional image is designed and completed.The components of the vision system are introduced both in hardware and software.The hand-eye calibration experiment of the camera is designed and completed by the principle of least squares calibration.Then the position information of the shaft is obtained by contour detection and shape matching algorithm and the angle information of the axis is obtained by the method of the minimum circumscribed rectangle.According to the position information and angle information of the shaft,the object recognition and positioning for the shaft of different poses is completed.In the end,the whole experimental process of the shaft-hole assembly based on the two-dimensional image is implemented.Finally,the algorithm is improved and the performance is evaluated.Using multiple templates instead of a single template improves the recognition rate of the algorithm.At the same time,using the contour area as a constraint condition reduces the number of matching contours and improves the recognition speed of the algorithm.In addition,it is verified that the algorithm has translation,rotation and scale invariance.