Research On Neural Network Intelligent Control Technology Of Robot Systems

The trajectory tracking control problem of robot system is an important research topic in the field of control. In actual work process, the control precision, reliability and dynamic performance of robotic systems are often affected by some unavoidable external interference,nonlinear, and uncertainty, etc. Thus, the research on trajectory tracking control problems of the robot control under complex environment is of great theoretical significance and has high application value. Due to its abilities in adaptation, neural network has been proven to be a powerful and effective method for controlling uncertain and complex nonlinear robot systems, associative memory, and nonlinear approximation. It is the key technology of this thesis. Motivated by the discussions above, this thesis studies the neural network control theory, and investigates the neural network intelligent control technology of robot systems.The main contributions of this thesis are summarized as follows:(1)Stability theory of neural network is investigated. Based on M matrix and Lyapunov stability theory, this thesis investigates the dynamical behavior of a class of time-delay TS fuzzy Hopfield neural network with discontinuous activation functions before obtains a globally asymptotically stability criterion of T-S fuzzy Hopfield neural network in its fixed point. In addition, a robust adaptive control method is applied for a class of discontinuous Cohen-Grossberg neural networks to ensure the stability on its equilibrium point. Finally,simulation experiments validate the effectiveness and controllability of our robust adaptive controller.(2)Neural network intelligent control technology is investigated. A robust Sliding Mode Control(SMC) method is proposed for a class of delayed T-S fuzzy neural network model.By Lyapunov-Krasovskii functional theorem and LMI methods, the sliding mode control method can achieve the tracking control of delayed T-S fuzzy neural network. Based on Lyapunov stability theory, the design of SMC is able to guarantee the robust asymptotic stability of delayed T-S fuzzy neural network. In addition, the tracking error converges asymptotically to zero. Through two numerical experiments the feasibility and effectiveness of the proposed controller are verified.(3)A robust adaptive control approach based on neural network is developed for a class of motor-driven robotic systems. First, the robust adaptive controller is designed based on Lyapunov stability theory to ensure the systems’ robustness and stability. Then, the Radial Basis Function(RBF) neural network is used to identify the unstructured system dynamics directly owing to its excellent compensation nonlinear function ability. Finally, a simulation experiment is carried out to test the adaptive performance and robustness of the obtained results.(4)Robot adaptive neural network fuzzy control method is proposed for a class of mobile robots. Based on robust adaptive control methods and Fuzzy Cerebellar Model Articulation Controller(FCMAC) neural network, which can accurately approximate unknown dynamics of mobile robotic systems, we realize the accurately tracking control for the mobile robotic system. The designed controller can guarantee the stability of the system by applying Lyapunov stability theory. Controllability and effectiveness of our results are verified by Matlab simulation.(5)A robust adaptive sliding mode control scheme is proposed for mobile robot manipulators in the presence of parametric uncertainties and external disturbances. An adaptive learning algorithm combined with the Fuzzy Gaussian Potential Function Neural Network(FGPFNN), which can accurately approximate unknown dynamics of systems, is an attractive control approach for mobile robots. According to the Lyapunov stability theory, it is shown that the proposed controller can guarantee the stability of the closed-loop system and require no prior knowledge about dynamics of the robot and off-line learning phases.Finally, the effectiveness of the proposed adaptive control approach is illustrated through comparative simulations on robot manipulator.