Research On Trajectory Planning And Motion Control Strategy Of Dual Arm Robot Based On Kinect

With the continuous progress of industrial technology,the application scenarios of robots are becoming more and more widespread.The deep integration of artificial intelligence,machine vision technology,and robotics technology enables robots to replace manual labor in an increasing number of situations and autonomously complete various repetitive and customized tasks in the production process.However,in certain special environments,such as space,deep sea,nuclear radiation and other dangerous work scenarios,due to the complexity and uncertainty of the on-site situation,work tasks that require real-time human participation in decision-making and implementation,thus robots need to perform various actions based on human instructions through human-robot interaction and control technology.This article proposes a trajectory planning and motion control strategy for a dual-arm robot based on Kinect,in response to the above application background.The on-site image is fed back to the human eye in real time through a camera,the human brain makes corresponding arm movements according to task requirements,the mechanical arm operates based on arm pose data and gestures,and the robot replaces humans to work in dangerous environments.The main research content of this article is as follows:Firstly,in response to the problem of difficult arm control due to the high noise in the raw motion data collected by Kinect,an adaptive Holt filter method is proposed.By combining adaptive parameters with the Holt filter method,the method achieves a smooth processing of random noise and peak noise in motion data,obtaining high-precision motion capture data.Secondly,to address the problem of difficulty in selecting the optimal trajectory due to the large number of motion trajectories generated during arm trajectory planning,the transformation relationship between pose and joint angles is studied through kinematic and singularity analysis.The working space of the robotic arm is calculated using the Monte Carlo method,and self-collision detection is established based on bounding box methods to avoid self-collision problems in trajectory planning of the dual arms.According to task requirements,the trajectory planning algorithms for joint space and Cartesian space are classified and thoroughly studied,and the correctness of the algorithm is verified through experiments,obtaining the optimal trajectory.Furthermore,the study proposes a control strategy based on human-machine mapping and trajectory tracking to address the problem of remote manipulation of robotic arms and low control precision.The strategy realizes human perception-based control of the robotic arm and gripper through camera calibration,human-machine pose mapping,and gesturegripper mapping.In addition,a fuzzy PID controller is designed to achieve precise trajectory tracking,thereby improving accuracy.Finally,an experimental platform is built for the application background of the article.Joint-space trajectory planning and Cartesian-space trajectory planning experiments of a Kinect-based dual-arm robot are designed for quick point-to-point movements of the robotic arm and grasping scenes.The effectiveness of the perception-based control method for the robotic arm is validated,the trajectory planning algorithm and motion control strategy are optimized,and the accuracy of controlling the end pose of the robotic arm is improved,and synchronisation of the gripper control,which lays the foundation for future engineering applications.