Six Degree-of-Freedom Dual-Arm Robot Online Teaching And Trajectory Planning Research

The traditional methods of teaching industrial robotic arms mainly involve direct programming or on-site manual teaching,but these methods have poor adaptability to different work scenarios and require the teacher to have a certain technical background.With the increasing complexity of work tasks and work scenarios,people hope to have more adaptable and easier-to-operate methods to interact with robotic arms.Therefore,based on this point,this thesis designs a real-time teaching system for a six-degree-offreedom industrial robotic arm based on the Kinect sensor.This system realizes the function of human-arm teaching of robotic arms,and in order to further ensure the stability and smoothness of the motion trajectory,the optimized particle swarm optimization algorithm based on differential evolution(IHPSO-DE)is used to optimize the teaching trajectory.The main research tasks are as follows:(1)Firstly,a dual-arm teaching experiment platform was designed,which includes the software and hardware design of the upper and lower computers,as well as the selection of various components.Then,the kinematics of the robot arm were calculated and verified.To further describe the velocity relationship between the joints and the end of the robot arm,the velocity Jacobian matrix was solved,and the singular configurations of the robot arm were analyzed based on this.Finally,the Monte Carlo method was used to calculate the workspace of the robot arm.(2)In order to better understand the Kinect data acquisition characteristics,an experimental analysis was first conducted on the factors that affect the accuracy and stability of Kinect acquisition.Subsequently,in order to eliminate the fluctuations in the collected data and improve the real-time performance of the teaching process,adaptive Kalman filtering was used to process the data.After converting the filtered wrist joint pose data to the mechanical arm motion space,its inverse solution can be obtained to obtain the rotational angles of each joint of the mechanical arm.In order to further improve the stability and accuracy of joint motion,a cubic spline curve was used to interpolate and smooth the angle values.(3)Due to the arm’s inherent tremor and the influence of Kinect’s data collection characteristics,the motion trajectory obtained directly through teaching is not smooth enough.Therefore,this paper uses a 3-5-3 polynomial composite function to replan the motion trajectory between key points.Furthermore,based on this,in order to further shorten the trajectory running time,an improved hybrid particle swarm optimization algorithm(IHPSO-DE)is used to optimize the trajectory running time.Finally,to demonstrate the superiority of this algorithm,it is compared with the standard PSO,standard DE,and CPSO algorithm-planned trajectories.Through experiments,it is proven that the motion time of the trajectory planned by the IHPSO-DE algorithm is reduced by 14.70%,13.98% and 12.56% compared to the above three algorithms,respectively.