Research On Neural Network Robust Control Of Robot With External Disturbance And Parameter Uncertainty

With the progress of science and technology,robots are widely used in various fields of human life,among which the research of robot trajectory tracking control is a very important aspect of robot technology.In practical application,the robot system will be disturbed by the external environment and the model parameters are uncertain.The robust control method has strong advantages in dealing with uncertainties such as robot system parameters and external disturbances,and the selection of robust terms is difficult in practical application.A larger robust term that overcompensates the system will consume too much energy,while a smaller robust term that undercompensates the system may cause instability.The main work of this thesis is as follows:A robust control strategy based on RBF neural network is proposed to solve the problems of uncertain model parameters and external interference.In order to obtain an accurate robot dynamic model,RBF neural network is used to identify model uncertainties,and robust control is used to reduce the influence of external interference and identification errors,so as to ensure the stability of the system.Because it is difficult to select the structural parameters in the robust control,the genetic algorithm is used for off-line optimization to get the optimal value.In order to avoid premature convergence,a local degenerate operator is introduced into the genetic algorithm selection strategy to increase population diversity.On the basis of the above research,considering the convergence of system state error in finite time,a sliding mode control strategy of RBF neural network is proposed.In order to guarantee the transient and steady state performance of the system,the tracking error is transformed by the system performance function,and the conditional constraint of the performance function is transformed into equivalent unconditional constraint.In order to reduce the chattering of the sliding mode and accelerate the convergence speed of the system,the non-singular terminal sliding mode surface was designed,and the structural parameters of the sliding mode surface were adjusted to reduce the time of the trajectory moving point reaching the sliding mode surface.The sliding mode interference observer is used to compensate the neural network modeling error and external interference to enhance the robustness of the system.Finally,the stability of the system is proved by Lyapunov theorem.The simulation results verify the rationality and effectiveness of the control strategy designed in this thesis,the comparison of simulation results verifies the effectiveness of the control strategy designed,through the inverted pendulum experiment verifies the practicability of the control strategy designed.