Trajectory Tracking And Vibration Suppression Of Flexible Robots Based On Neural Networks

With the rapid development of robotics technology,high-speed and lightweight flexible robots are becoming a hot research topic for international scholars.The flexible components of robots are mainly reflected in the flexibility of the arm and joint,and the vibration generated during the working process can seriously affect the tracking accuracy of the flexible robot.Therefore,how to suppress elastic vibration while achieving high-precision trajectory tracking is a difficult problem for flexible robot control.In response to the above issues,considering that flexible robots are highly nonlinear coupled systems,which makes it difficult to obtain accurate models for the robot system,a vibration suppression control strategy based on neural networks is proposed.Firstly,the dynamic equations of the flexible joint robot were derived using Spong’s simplified spring model of flexible joints and the Lagrange method.The dynamic equations of flexible arm robots are derived using the assumed modal method and Lagrange method for flexible arm deformation,providing a design basis for subsequent controller derivation of flexible robots.Secondly,aiming at the problem of trajectory tracking and vibration suppression of flexible joint robot with limited output,a neural network adaptive controller based on singular perturbation theory is designed for logarithmic obstacle Lyapunov function.In order to ensure that the tracking error is within the preset range,a specified performance function and joint flexibility compensator are introduced.A neural network adaptive controller based on the specified performance is designed,and the effectiveness of the two algorithms is verified through simulation experiments.Furthermore,for the trajectory tracking and vibration suppression of flexible manipulators,a sliding mode controller with neural network adaptive compensation is designed based on singular perturbation theory.In order to compensate the approximation error of neural network and eliminate the "chattering" problem caused by sliding mode variable structure,fractional sliding mode function and hyperbolic tangent function are introduced,and neural network adaptive controller based on fractional sliding mode is designed,and the effectiveness of the two algorithms was verified through simulation experiments.Finally,a physical experimental platform scheme was designed,using the STM32 microcontroller as the lower computer,the PC as the upper computer,and the JY61 P sensor to detect vibration data.The RP(Rotation Parallel)flexible arm robot experimental platform was built.Considering the real-time and dynamic characteristics of engineering applications,a PID vibration suppression controller was designed,and a vibration characteristic acquisition program and human-machine monitoring interface were designed based on Lab VIEW.The vibration characteristics of the flexible arm were analyzed.