Research On Friction Compensation Strategy Of Industrial Robot Joint

With the rapid advancement of industrial modernization,the development of China's manufacturing industry has also taken a new step,industrial robots are an important part of today's manufacturing industry,they are used in a variety of production scenarios,and can replace manual handling,welding and other tasks.However,in engineering practice,it is common for industrial robots to operate with low accuracy,which is mainly affected by the friction force of the robot joints.Aiming at improving the accuracy of industrial robot,this paper makes a deep research on the joint friction compensation of industrial robot.Aiming at the problem of joint friction compensation for industrial robots,this paper summarizes the friction phenomena and friction models,and then designs a joint friction compensation strategy based on tribeck friction model.Firstly,the measurement methods of friction moment of robot joints in simulation environment and physical environment are introduced.Then genetic algorithm is introduced,and the parameters of Stribeck friction model are identified by genetic algorithm.Finally,simulation experiments and real experiments are designed to verify the effectiveness of the joint friction compensation strategy based on tribeck friction model.In this paper,the Simulink tool in MATLAB is used to carry out simulation experiments,the simulation results show that the friction compensation strategy based on tribeck friction model can effectively improve the accuracy of the robot joint,which is manifested in the fact that the speed and position tracking errors of the robot joint become smaller.At the same time,a six-axis industrial robot experiment platform is carried out in this paper,the experimental results show that the robot joint friction compensation strategy based on tribeck friction model can effectively reduce the position tracking error of the robot joint in practical application.In this paper we use Reinforcement Learning technology to solve the problem of joint friction compensation of industrial robots,designed the robot joint friction compensation strategy based on strategy gradient algorithm.Firstly,we demonstrated the feasibility of applying Reinforcement Learning technology to joint friction compensation of industrial robots,defined the key elements of Reinforcement Learning for joint friction compensation of industrial robots.And then introduced the prior knowledge,simulation experiments show that the introduction of prior knowledge is helpful to enhance the learning efficiency of Reinforcement Learning.Finally,the simulation experiment of robot joint friction compensation based on strategy gradient algorithm is designed by using MATLAB and Python hybrid programming,at the same time,the experiment is carried out on the six-axis industrial robot platform,and compared with Stribeck friction model without friction compensation and robot joint friction compensation strategy.The experimental results show that the friction compensation strategy based on the strategy gradient algorithm can effectively compensate the friction produced by the robot joints in both simulation and physical environments,and it is superior to the friction compensation strategy based on non-friction compensation and Stribeck friction model.