Research On Robot Drag Teaching And Trajectory Optimization Technology Based On Predictive Control

With the introduction of "Made in China 2025" and "Industry 4.0",the collaboration and interactivity of industrial robots have become new research hotspots,and teaching as a prerequisite for the application of industrial robots has also been put forward new requirements.The traditional teaching methods of industrial robots are mainly based on teaching by the teaching device and software.The teaching operation of the teach pendant is cumbersome and the professional requirements of the operators are high and it cannot be used efficiently in workplaces with complex trajectories such as curves or curved surfaces.The software teaching is separated from the robot body,and the real-time teaching cannot be guaranteed.In this context,drag teaching has been widely studied due to its simple,efficient operation and low requirements on the operator.Based on the comprehensive comparison of the existing drag teaching methods and the robot's own characteristics,and based on the idea of human-robot interaction,this paper finally conducts the research on robot drag teaching and trajectory optimization technology based on predictive control.The main research contents of this paper include:(1)Research on the Hardware and Control System of the Drag Teaching Robot.Build a hardware platform and control system according to the technical characteristics of robot dragging and teaching,Taking a conventional six-degree-of-freedom robot as the operating body,Establish an open control system platform based on PC + motion control card,Take GOOGOLTECH GTS-800-PV(G)-PCI eight-axis motion control card as the core of the control system;The daisy chain RS232 communication protocol is adopted between the controller and the servo drive to ensure the real-time communication;The software platform is based on Simulink under MATLAB2012 to design the control program module.(2)Research on Drag Teaching Model.Analyze the dynamics of the robot,and establish the theoretical dynamics model of the robot through the Lagrange method.At the same time,the robot friction model is established by Coulomb friction + viscous friction,At the same time,according to the established dynamics,the external force is predicted and estimated based on the generalized momentum,and the direction and magnitude of the operating force are determined to compensate for the static friction force.According to the established drag teaching model,the parameters to be identified are determined and then the parameters are identified.Aiming at the shortcomings of traditional identification methods,intelligent algorithms are applied to parameter identification.It is proposed to optimize BP neural network algorithm identification based on genetic algorithm.Through genetic algorithm to optimize parameters such as weights and thresholds in BP neural network,the identification accuracy can be improved at the same time.Overcome the defects of conventional BP neural network algorithms that are easy to fall into local convergence and long operation time.Then establish the excitation trajectory of the Fourier series,collect sample data for training and identification,and identify relevant parameters.(3)Research on Trajectory Tracking and Reappearance.According to the drag teaching model and the identified dynamic parameters,Establish a teaching program in the MATLAB2012 software platform to complete dragging and teaching;After the teaching is completed,according to the saved data,First,smooth the teaching trajectory of the data through five-point three-time filtering.Then a dynamic feedforward PID control trajectory tracking program is established for trajectory tracking.