With the construction of a new manufacturing system with intelligent manufacturing as the core,industrial robots are playing a more and more important role in human industrial production.The intelligent requirements of modern industrial production for robots are also increasing day by day.Although many scholars have explored and studied the obstacle avoidance planning of robots,there are still many problems to be solved.Combined with the increasingly perfect machine vision technology and based on the binocular stereo matching principle,the following aspects of the dynamic obstacle avoidance of the robot are studied in this paper:1.The first chapter expounds the research background and significance of the subject,and makes a certain investigation and study on the development history and development process at home and abroad of industrial robot technology,machine vision technology and path planning technology.2.The second chapter reconstructs the working environment of the robot based on the principle of binocular vision.The depth map of the robot’s workspace is obtained by using the binocular stereo matching algorithm with polar constraint,and the workspace is reconstructed.3.The third chapter simply analyzes the existing strengths and weaknesses of collision detection algorithm in the three-dimensional environment.The application principle of A-star algorithm in robot path planning is analyzed,and the iteration efficiency of the algorithm is improved by optimizing the storage mode of nodes.The dynamic weight coefficient is introduced to reduce the collision probability between the planned path and obstacles.A dynamic obstacle avoidance strategy based on A-star algorithm is proposed.4.The fourth chapter aims at the problems of deviation from optimal solution,lack of purpose and uneven path in fast random spread tree(RRT)algorithm,combining RRT algorithm with A-star algorithm and introducing heuristic function to guide the expansion tree to rapidly expand towards the target node.In view of the local minimum trap existing in heuristic function,regression constraints are added to enhance the ability of the algorithm to explore the unknown space.The dynamic obstacle avoidance strategy of caching initial tree,pruning and renaturalization is adopted to avoid moving obstacles.Bessel fitting is applied to the final path to further reduce the jitter of the robot.5.The last chapter simply analyzes the requirements and objectives of the robot three-dimensional motion obstacle avoidance system based on machine vision,and builds the hardware and software platform.The hardware platform includes the machine vision platform and the robot experiment platform;the software platform writes the upper GUI control interface in MATLAB.Finally,the feasibility of the proposed method is verified on a real platform. |