Modeling And Neural Network Control Of Flexible Systems

The flexible system covers many different objects such as flexible robotic manipulators,bionic flexible flapping wing aircraft and flexible buildings.With a large number of applications of flexible systems,its control theory and method issues have become a prospective high-tech research direction,which attracts concerns from both academic and industrial fields.At present,the control theory and method of flexible systems,such as the tracking and vibration control of multi-link flexible manipulators,the constraint control of flexible buildings under natural disasters,and the fault-tolerant control of bionic flexible flapping-wing robots,has developed into a common scientific problem,which is extremely challenging.In order to solve the technical problems of modeling and intelligent control of uncertain flexible systems with environmental adaptability,the thesis makes a systematic and detailed study on modeling mechanism and control strategy of several flexible systems.Firstly,the dynamic model of the flexible robotic manipulator is established using assumed mode method,which overcomes the challenge from the system dynamics being infinite dimensional.The fuzzy neural networks with uniform approximation performance are designed to solve the system dynamic uncertainties.A neural network control strategy based on a high-gain observer is proposed to estimate the immeasurable states in practice.The extensive experiments have been carried out on the Quanser platforms to verify the effectiveness of the proposed methods.Secondly,considering the flexible building systems with output constraints,an adaptive reinforcement learning control scheme is designed based on the Actor-Critic algorithm,while an auxiliary system and a disturbance observer are proposed based on barrier Lyapunov function.The developed approach addresses the output constraints and vibration suppression issues of the flexible buildings under unexpected natural disaters.The experimental verification has been successfully implemented on the Quanser platforms.These results break through the limitations of traditional control methods that cannot deal with distributed disturbances,infinite dimensionality and system uncertainties.Thirdly,the dynamic model of the flexible flapping-wing aircraft is established by an improved rigid finite element(IRFE)method.A novel adaptive fault-tolerant controller based on the fuzzy neural network and nonsingular fast terminal slidingmode control scheme are proposed for tracking control and vibration suppression of the flexible wings,while successfully addressing the issues of system uncertainties and actuator failures.Co-simulations through MapleSim and MATLAB/Simulink have been carried out to verify the performance of the proposed controller.This thesis analyzes the dynamic characteristics of several flexible systems,and studies the vibration control and optimization problems in the real engineering.The research results will provide new solutions for the modeling mechanism and control design of flexible systems,and further contribute to the development of the mechanical design and the control theory for flexible systems.