Research On Motion Planning Of 6-DOF Robot Arm Based On Multi-Kinect Vision

This paper developed under the research fund for the doctoral program of higher education named “the Bilateral Servo Control Strategy of tele-operation of electro-hydraulic underactuated redundant robot”(No.20130061110009).Taking the 6-DOF manipulator as the research object,the motion planning based on the Multi-Kinect acquisition is used to study the motion.In this paper,a multi-visual sensor arrangement is proposed for the working space of the 6-DOF manipulator.Based on the arranged Kinect sensor,it is calibrated to obtain the conversion relationship between multiple Kinect and internal and external parameters.According to the characteristics of Kinect vision sensor,BP neural network algorithm is used to realize data fusion between multiple sensors.The obstacle information collected by Kinect is filtered and real-time trained to obtain the mapping relationship between multiple Kinect.The fusion results show that this method can meet the requirements of the fusion accuracy of Kinect camera.The obtained point cloud image is denoised,and finally an octree map of the surrounding environment of the robot is established according to the processed point cloud data.Aiming at the problem of six-degree-of-freedom manipulator motion planning,an improved RRT algorithm is proposed.Based on the RRT algorithm,the greedy idea of Dijkstra algorithm is introduced,and the RRT algorithm is combined with the Dijkstra algorithm.Establish a threedimensional visualization obstacle environment model,through the comparison of RRT,RRT*,BRRT and improved RRT in the simulation,The improved RRT algorithm speeds up the expansion of the random tree in the local area,reduces the blindness and increases the efficiency of the algorithm.The robot end path is more reasonable than the path generated before the improvement.In the ROS simulation environment,the improved RRT algorithm is tested by combining the octree map generated by Kinect sensor.Finally,the improved RRT algorithm is transplanted into the robot arm of the laboratory to verify the feasibility of the algorithm.The above simulation and experimental results show that the vision-based six-free robot motion planning obtains a trajectory with no obstacle path and the least path cost.The trajectory of the robot arm is smooth and the smooth planned path meets the expected result.