Adaptive Dynamic Surface Control Strategy For Several Classes Of Strict Feedback Nonlinear Systems And Its Application

In the control problem of practical system,because of the different physical characteristics of the controlled object,the construction of its mathematical model is also different.For the control of nonlinear systems,the existence of modeling errors,measurement noise,model simplification,nonlinear input links,communication rates and other factors will have a great impact on the performance of the control.Therefore,different control schemes are often used to deal with different control models,which is also the reason why the control schemes of nonlinear systems are difficult to be unified.In fact,strict feedback nonlinear systems are a common class of nonlinear systems,and many practical systems can eventually be transformed into strict feedback nonlinear systems.Although there have been a lot of research results on strict feedback nonlinear system control,most of them have verified the control schemes by means of simulation or hardware in the loop simulation,and have not really been put into practical application.In this dissertation,several classes of strict feedback nonlinear systems,including a class of strict feedback nonlinear systems with unknown system state,a class of discrete-time strict feedback nonlinear systems,a class of strict feedback nonlinear systems with butterfly hysteresis input and a class of large-scale strict feedback nonlinear time-delay systems with hysteresis input,are studied.Based on the adaptive dynamic surface control algorithm,fuzzy logic,neural network approximation,high gain filter hysteresis quantizer,finite cover lemma and implicit inverse algorithm are used to realize the accurate control of nonlinear system.The proposed control scheme is applied to the actual physical system,and the effectiveness of the control scheme is verified by building the corresponding experimental platform.The specific research contents are as follows:(1)For a class of strict feedback nonlinear systems with unknown system state,an adaptive dynamic surface quantization control scheme based on fuzzy approximator is proposed to solve the problem that the system state is not measurable in the control system.Using high gain K filter,a state observer is designed to estimate the unknown state of the system.The hysteresis quantizer is introduced to effectively alleviate the limitation of communication channel bandwidth and data transmission rate on the control effect in the process of signal transmission,avoid the shake in the process of quantization,and improve the control efficiency.The fuzzy approximator is used to estimate the unknown nonlinear function in the control system on-line,and an adaptive dynamic surface control scheme is designed,which overcomes the"differential explosion"problem in the traditional backstepping method,and realizes the L_?performance of tracking error through initialization technology.The proposed control scheme is applied to the flight control of UAV.The effectiveness of the proposed control scheme is verified by building UAV control experimental platform and comparing it with PID-LQR method and backstepping-sliding mode method.(2)For a class of discrete-time strict feedback nonlinear systems,a discrete-time adaptive dynamic surface tracking control scheme is proposed,which solves the problem that the traditional continuous time control signal can not be directly applied to digital hardware.Different from continuous time control scheme,discrete time control is more in line with the operation mode of computer and network control.The radial basis function neural network approximator is used to estimate the unknown nonlinear function relationship in the discrete-time system,and the digital first-order low-pass filter is introduced to predict the future virtual control signals,which avoids the model transformation problem of the discrete-time nonlinear system.Finally,the proposed control scheme is applied to the flight control of UAV,and compared with the discrete-time backstepping method.The control effect is verified by the experimental platform.(3)For a class of strict feedback nonlinear systems with butterfly hysteresis input,an adaptive implicit inverse control scheme based on neural network is proposed,which solves the influence of butterfly hysteresis nonlinear input on the control system for the first time.A new butterfly hysteresis model is established to predict the butterfly hysteresis effect in control input.An implicit inverse control scheme is designed to eliminate the butterfly hysteresis effect,which avoids the construction of the direct inverse model of hysteresis.Combined with neural network technology,the implicit inverse method is applied to the design of output feedback control scheme.The L_?norm with arbitrarily small tracking error is realized while eliminating butterfly hysteresis.The proposed control scheme is applied to the flexible intelligent material drive motion system.The advantages of the proposed control scheme are verified by building a dielectric elastomer drive control experimental platform and comparing it with the backstepping method and the proposed control scheme without considering hysteresis.(4)For a class of large-scale strict feedback nonlinear time-delay systems with hysteretic inputs,a decentralized adaptive dynamic surface implicit inverse control scheme based on fuzzy approximator is proposed to solve the influence of subsystem interaction and multi-hysteresis nonlinear input links on the control of large-scale systems.The design of implicit inverse compensator can effectively alleviate the phenomenon of multiple hysteresis loops in large-scale systems.At the same time,the design of implicit inverse compensator also replaces the traditional hysteresis inverse model which is difficult to construct.A search mechanism for obtaining the actual control signal from the hysteresis temporary control law is proposed.The fuzzy logic system and the finite cover lemma are used to deal with the time delay and unknown nonlinear coupling relationship in the system,and the L_?performance with arbitrarily small tracking error is obtained by using the initialization technique.The designed control scheme is applied to the giant magnetostrictive drive system.The effectiveness of the proposed control scheme is verified by building a three-axis giant magnetostrictive motion control experimental platform and comparing it with PID method and the proposed control scheme without considering hysteresis method.