【作者】 沈启坤；

【导师】 张天平；

【作者基本信息】 扬州大学 ， 计算机应用技术， 2007， 硕士

【摘要】 对于具有执行机构非线性特性的不确定非线性系统的自适应控制研究，是智能控制领域的研究热点之一。本文就此领域的相关问题展开系列研究。第一，针对一类非仿射非线性系统，利用隐函数定理和泰勒公式，将系统由非仿射型转变为仿射型，然后基于变结构控制原理，并运用模糊逻辑系统对未知函数进行在线逼近，提出了一种具有监督控制器的自适应模糊控制方法。该方法在考虑到外界干扰的情况下，通过监督控制器保证闭环系统所有信号有界，进一步通过引入最优逼近误差的自适应补偿项来消除建模误差的影响。通过Lyapunov方法，证明了跟踪误差收敛到零。第二，首先针对一类具有未知类反斜线回滞的非线性SISO系统，基于变结构控制原理，利用模糊逻辑系统对未知函数进行在线逼近，提出了具有监督控制器的自适应模糊控制器设计方案。该方案首先通过监督控制器保证闭环系统的稳定性，由此确定出建模的有界区域，然后再通过引入最优逼近误差的自适应补偿项，保证跟踪误差收敛到零的一个邻域。其次，将上述结果推广到非线性MIMO耦合系统，提出了一种分散自适应模糊控制方案。对每一个子系统，仅通过子系统的自身信息在线调节模糊系统的可调参数、增益，从而实现了对动态不确定性及建模误差进行自适应补偿。通过李亚普诺夫方法，证明了跟踪误差收敛到零的一个邻域内。第三，针对具有未知死区模型的两种结构不同的系统，分别提出了鲁棒自适应模糊控制方案。首先，针对一类严格反馈的SISO时滞非线性系统，基于模糊系统的逼近能力、后推设计方法和修改的积分型李亚普诺夫函数，运用简化死区模型及Nussbaum函数的性质，提出了一种自适应控制方案。该方法不需要控制增益符号的先验知识且取消了和死区参数已知的条件。此外，该方法通过引入逼近误差的自适应补偿项来消除建模误差和参数估计误差的影响。理论分析证明了闭环系统分别是半全局一致终结有界。其次，研究了一类具有死区的MIMO非线性系统的稳定跟踪问题，提出了相应的具有监督控制器的控制方案，并从理论证明了闭环系统分别是全局一致终结有界。第四，针对具有未知非线性死区输入的线性／非线性系统，分别提出了相应的模糊控制方案。首先，针对一类具有未知非线性死区的SISO非线性系统，基于滑模控制原理，利用模糊系统的逼近能力，提出了一种自适应模糊控制方案。该方法利用一种简化非线性死区输入模型，取消了非线性死区输入模型的倾斜度相等以及死区边界对称的条件。此外，该方法还取消了非线性死区输入模型各种参数已知的条件。该方案通过引入逼近误差的自适应补偿项来消除建模误差和参数估计误差的影响。通过李亚普诺夫综合方法，证明了闭环系统半全局一致终结有界，跟踪误差收敛到零。其次，研究了一类线性时滞大系统的稳定控制问题，提出了相应的自适应控制方案。最后，研究了一类具有状态时滞和未知非线性死区输入时滞的不确定系统的模糊控制问题，利用并行分配补偿的思想，提出了一种状态反馈模糊控制器，并基于线性不等式(LMI)，通过构造合适的Lyapunov-Krasovskii函数，给出了状态反馈模糊控制增益存在的充分条件。理论分析证明了闭环系统是渐近稳定的。第五，研究了一类具有死区饱和输入的不确定时滞混沌系统的跟踪问题。基于滑模控制原理，提出了一种自适应控制的新方案，并从理论上分析证明了该控制方案的有效性。通过本文的研究，较好的解决了一些具有执行机构非线性特性的不确定非线性系统的模糊控制问题。仿真表明所提控制方案的有效性。更多还原

【Abstract】 In recent years, adaptive control for uncertain systems with actuator nonlinearityusing fuzzy systems has been a hot topic in the field of intelligent control. Somecorrelative problems in this area are studied in this paper.Firstly, using Implicit Function Theorem and the Talor series expansion, the form ofthe non-affine nonlinear systems is transformed to the form of affine nonlinear systems.Based on the principle of variable structure control, an adaptive fuzzy control strategywith supervisory controller is designed for the transformed nonlinear system. With thehelp of a supervisory controller, the reSulting closed-loop system is globally stable inthe sense that all signals involved are uniformly bounded. Furthermore, the adaptivecompensation term of the optimal approximation errors is adopted to reduce the effectsof modeling error. By theoretical analysis, it is shown that tracking errors converges tozero.Secondly, based on the principle of variable structure control and the capability ofthe fuzzy systems with linear adjustable parameters, an adaptive fuzzy control strategywith supervisory controller for a class of uncertain nonlinear systems with unknownbacklash-like hystersis is developed. With the help of a supervisory controller, theresulting closed-loop system is globally stable in the sense that all signals involved areuniformly bounded. Furthermore, the adaptive compensation term of the optimalapproximation error is adopted to reduce the effects of modeling error. By theoreticalanalysis, it is shown that tracking error converges to a small neighborhood of zero. Next, the above result is extended to the MIMO interconnected nonlinear systems withunknown backlash-like hysteresis, a decentralized adaptive fuzzy control strategy isdeveloped. The plant dynamic uncertainty and modeling errors are adaptivelycompensated by adjusted the parameters and gains on-line for each subsystems usingonly local information. By theoretical analysis, it is shown that the closed-loop fuzzycontrol systems are globally stable, with tracking error converging to a neighborhood ofzero.Thirdly, two design schemes are proposed for two class of systems with unknowndead-zone, which their structures are different, respectively. Begin with, based on theapproximation capability of fuzzy systems and backstepping technique, and byintroducing modified integral-Lyapunov function, and by utilizing the simplified deadzone model and the property of Nussbaum function, a design scheme of adaptive fuzzycontroller for a class of nonlinear time-varying time-delayed systems with unknownfunction control gain, is proposed. The approach does not require a priori knowledge ofthe sign of the control gain and removes the conditions that dead-zone modelparameters are known constants. In addition, the adaptive compensation term of theapproximation error is adopted to minify the influence of modeling errors and parameterestimation errors. By theoretical analysis, the closed-loop control system is proved to be semi-globally uniformly, ultimately bounded. Then, the problem of adaptive fuzzycontrol for a class of MIMO time-delayed nonlinear systems with unknown dead-zoneis discussed, and a design scheme of adaptive fuzzy controller with a supervisorycontroller is proposed. By theoretical analysis the resulting closed-loop system, isglobally uniformly ultimately bounded with tracking errors converging to zero.Fourthly, different design schemes for a class of linear or nonlinear systems areproposed, respectively. First, based on the principle of sliding mode control, using thesimplified nonlinear dead-zone model and the fuzzy systems with linear adjustableparameters which are used to approximate plant unknown functions, a design scheme ofadaptive fuzzy controller for a class of SISO nonlinear systems is proposed. Theapproach removes the conditions that the dead-zone slopes and. boundaries are equal, symmetry, respectively. In addition, it does not require the assumptions that allnonlinear dead-zone model parameters are known constants. The adaptive compensationterm of the approximation errors is adopted to minify the influence of modeling errorsand parameter estimation errors. By theoretical analysis, the closed-loop control systemis proved to be semi-globally uniformly ultimately bounded with tracking errorsconverging to zero. Then, the problem of adaptive fuzzy control for a class oflarge-scale time-delayed systems with unknown nonlinear dead-zone is discussed, and adesign scheme of adaptive fuzzy controller is proposed. Lastly, the concept of parralldistributed compensation(PDC) is employed to design the state feedback controller for aclass of uncertain time-delayed systems with input delay and unknown nonlineardead-zone. Based on LMI, by constructing appropriate Lyapunov-Krasovskiifunctional(LKF), sufficient conditions for the existence of fuzzy state feedback gain arederived. By theoretical analysis, the closed-loop control system is proved to be globallyasymptotically stableFinally, the chaos tracking control problem is studied for a class of uncertaintime-delayed chaotic systems with dead-zone and saturating input. Based on theprinciple of sliding mode control (SMC), a novel adaptive design scheme is proposed.The approach not only overcomes the drawback in the existing literature, but alsoremoves the assumptions that the dead-zone is symmetric and the slopes are known. Bytheoretical analysis, it is shown that tracking error asymptotically converges to zero.Through the research in this paper, some fuzzy, control problems for uncertainnonlinear systems with actuator nonlinearity have been properly solved. Numericalsimulation experiments of these Control schemes demonstrate their effectiveness.更多还原