Research On Robot Hub Grinding Systemdesign And Trajectory Planning Strategy

As an important component of automobiles,hub demand has continued to increase with the popularity of cars.Traditional manual polishing causes great harm to the human body and the quality of polishing is related to the experience of the worker,resulting in low efficiency and high cost.Although using industrial robots taught through demonstration can improve the polishing efficiency to some extent,it requires re-teaching when the production line changes models,which is very time-consuming.Although offline programming can save teaching time,there may be installation errors between the offline programming environment and the actual polishing environment,leading to polishing errors.This dissertation proposes a method of combining point cloud registration with offline programming to achieve automated polishing in response to the demand for hub polishing.First,the collected point cloud data of the hub is processed and the point cloud data used for registration is extracted.The hub point cloud data is projected and transformed into image data,and rough point cloud registration is completed by combining image processing.The hub pose data is obtained through ICP fine registration.Finally,the offline programming trajectory is optimized by the trajectory algorithm to make the trajectory smoother.The offline programming trajectory is converted into a real polishing trajectory using the transformation matrix obtained from registration,thus avoiding errors between the offline environment and the actual environment.The main content of this dissertation includes:Firstly,point cloud preprocessing algorithms such as pass-through filtering and radius filtering are used to optimize the point cloud data.When extracting the wheel hub point cloud data,a Z-axis layered traversal method is proposed to extract the top layer of wheel hub point cloud data.The original point cloud data is corrected by fitting the wheel hub plane data and re-establishing the coordinate system,reducing installation errors and making the data more accurate.Subsequent processing data is then extracted using an adaptive thresholding approach.Then,estimating the pose of the wheel hub based on a combination of point cloud and image.Firstly,the uppermost layer of the wheel hub point cloud data is projected onto a plane,then the projected data is transformed into image data using Kd-tree nearest neighbor search.Through methods such as dilation and contour detection,the key points needed for coarse registration are found,and the wheel hub point cloud data is rotated by the angle between the key points and the connection line between the key points and the coordinate system to complete the coarse point cloud registration.Finally,the wheel hub accurate pose is obtained through ICP fine registration.Finally,for the trajectory planning strategy,in order to solve the problems of uneven trajectory and discontinuous velocity that may exist in the path generated by offline programming software,an offline trajectory optimization program was developed and trajectory planning algorithms were used to optimize the trajectory to make the path smooth and reliable during polishing.