Research On Obstacle Avoidance Path Planning Method Of 6-DOF Manipulator Based On Visual Guidance

The industrial 4.0 era represented by intelligent manufacturing is coming.Robots are used in the manufacture of equipment,aerospace,liquids and logistics technology plays an important role in various sectors.Manipulator control technology is the core of robot technology important units.At present,in the scene of envelope space,complex environment and high dynamic,it is increasingly important to carry out the research on efficient obstacle avoidance path planning technology for robotic manipulators.In this paper,the 6-DOF manipulator is taken as the research object,and the key technologies for autonomous obstacle avoidance and path planning of the manipulator in dynamic complex scenes are studied with the help of machine vision.The main work and characteristics are as follows:Firstly,the image intelligent recognition technology was used to extract the features of the captured object,and two image processing methods are used for positioning: The contour based target detection algorithm obtains the screening results through template feature point matching? due to the traditional visual detection method,the target is easily affected by environmental factors such as uneven illumination and sensitive scale change.In order to solve the problems of limited computing power and high power consumption in the embedded environment of manipulator,an improved lightweight Yolov3 network model was proposed,and the effectiveness and accuracy of the network model were verified by training and testing the public data set and partial self-collected scene data.Then,the traditional RRT * path planning algorithm has strong randomness in sampling,and its search efficiency is low cost and long planning time,a new improvement strategy is proposed.E-greedy strategy is adopted to determine the weight.The selection between the old tree and the new sampling node is used to balance the sampling process with the combination of greed and randomness? using the ellipse sampling set,the sampling space is reduced to an ellipse area to speed up the search for better planning route.The superiority of the algorithm is fully verified by comparing the experimental results.Finally,in order to verify the validity of the above＿method in the obstacle avoidance planning of the mechanical arm,based on the ROS development environment,this paper builds a simulation model of the obstacle avoidance path planning of the sleeve mechanical arm using the modular design,and preliminarily proves the effectiveness of the algorithm.The improved algorithm is validated on the platform of the mechanical arm.The results show that the proposed algorithm makes the mechanical arm avoid obstacles effectively and safely during the grabbing process to complete the master planning.