Adaptive Neural Network Control Of Two-DOF Robotic Arm System With Output Constraint

Robotic Arm system is a multivariable complex nonlinear time-varying system with high coupling and many uncertainties.Because of the existence of these non-linear characteristics,it is very difficult to accurately model and control the robotic arm system.In this work,a Lagrange dynamic model is built for a two-degree-of-freedom robotic arm system driven by an electro-hydraulic servo system.The key methods of obstacle Lyapunov function,error conversion function,back-stepping control method and radial basis function neural network are combined to construct a controller with good performance,constraints on the angular displacement output of the system and the estimation of uncertainties in the model of the robotic arm system are realized step by step.Firstly,the theoretical knowledge of Lagrange equation,system stability theory,backstepping control theory,radial basis function neural network,obstacle Lyapunov function and error conversion function are introduced in detail.Then the dynamic model of robotic arm system is built based on Lagrange equation,and the system is controlled by PID parameter adjustment.The experiment proves that PID control is effective.The controller can track the desired output trajectory curve effectively without any external disturbance and output constraints.The tracking error can converge to a very small interval.Secondly,in order to restrict the output of the robotic arm system,the controller is constructed by using obstacle Lyapunov function combined with nonlinear backstepping control method and error conversion function combined with nonlinear backstepping control method,and the control effect is compared.It can be concluded that the control effect of combining angular displacement error conversion function with nonlinear backstepping control method is better,the bounds of constraints can be changed in real time.Finally,considering the uncertainties in the mathematical model of the robotic arm system,the RBF neural network control method is used to estimate and compensate the uncertainties.The controller is designed by using obstacle Lyapunov function combined with adaptive neural network control method and error conversion function combined with adaptive neural network control method.It can be proved that the RBF neural network control method is effective.The system can make the tracking error and the neural network estimation error reach the expected accuracy in a short time,and can accurately approximate the uncertainties in the model of the robotic arm system,which can prove that the control scheme designed in this paper is feasible and effective.