Visual obstacle avoidance method of manipulator based on ROS

Aiming at the problem that most industrial robots are unable to avoid obstacles independently in the working environment, a robot arm obstacle avoidance method based on ROS (robot operating platform) and guided by Kinect camera is proposed. Firstly, the robot model is established by using URDF file in ROS, and the kinematics configuration of the model is carried out; After that, the principle of camera calibration is analyzed, the internal and external parameters of the camera are calibrated, and then the hand eye matrix of the robot and the camera is obtained and the error is verified to provide support for visual obstacle avoidance; Finally, the physical simulation environment gazebo and Rviz are built for joint simulation. Obstacles are placed in gazebo. Kinect camera converts the obstacle point cloud data into Octomap map in real time and outputs it to Rviz visualization platform. The running program can see that the visualization platform controls the robot arm in the physical environment to carry out obstacle avoidance planning. The experimental results show that the robot arm guided by Kinect camera based on ROS platform can accurately carry out obstacle avoidance planning, which has reference significance for the application research of robot under ROS.