Neural Network Robust Control And Simulation Of Two-joint Robot

In recent years, with the development of science and technology, industrial robot has achieved significant advance. Multi-jointed robot has been used not only in a wide range of applications in the industrial field, but also more and more penetrated into other areas. While the most important, the core part of the robot is its control system, which directly affects the whole performance of the robot and the advanced degree. From the perspective of control theory, robotic systems have severe nonlinear characteristics, and it is difficult to obtain accurate robot mathematical model, couple with the disturbance of outside world and the model itself is not uncertainties, thus it will be reduced the effectiveness of multi-joint robot arm’s motion control.The article is firstly given introduces of the two-joint robot robust control and adaptive fuzzy inversion control respectively, and made a task of control design and simulation study. The experimental result can be seen robust control to ensure system stability, but the two-joint manipulator trajectory tracking is difficult to obtain good transient performance. Simultaneously, it can be improve the dynamic characteristics of two-joint manipulator position tracking by the adaptive fuzzy inversion control, but hard to get good result for the velocity and acceleration of manipulator trajectory tracking control.In view of it is harder to get the two-joint robot trajectory with good transient performance by the robust control and adaptive fuzzy inversion control respectively, this paper uses neural network robust control trajectory given expected value, with online identification by the neural network learning ability and real-time approximation the uncertainty of the whole control system, and with network learning algorithms and design of robust controller based on Lyapunov stability theory, and finally to eliminate the network approximation error by the robust controller. It is difficult to assume and obtain the value of the upper bound of uncertainty when the uncertainty of the system exceeded expectations hypothesis by the robust control, the robust control system will be divergent, even collapse. Therefore, the robustness of the neural network control can avoid complex calculations, and greatly improve the transient performance of the system, and conduct better for velocity and acceleration of the trajectory tracking control of two-joint robot position.