Font Size: a A A

Study On Robust Adaptive Control Based On Fuzzy Modeling And Observer

Posted on:2005-06-11Degree:MasterType:Thesis
Country:ChinaCandidate:H J GuFull Text:PDF
GTID:2168360125952708Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
As an important branch in the field of intelligent control, fuzzy modeling based robust adaptive control for nonlinear systems has been received more and more attention in recent years. Some correlative issues in this area are studied in this paper, such as the controller and observer design problem of uncertain nonlinear systems, which have the form of single-input-single-output (SISO) or multiple-input-multiple-output (MIMO). The design and analysis procedure is based on a series of control theories, which include Lyapunov stability theory, adaptive control theory, decentralized control theory, sliding mode control theory, fuzzy approximate theory, and so on. The main work in this paper is summarized as follows.Firstly, based on supervisory control strategy, a new design scheme of direct adaptive fuzzy controller for a class of SISO uncertain nonlinear dynamical systems with unknown function control gains is proposed. The adaptive compensation term of the optimal approximation error and a new robust term are adopted to minify the influence of the fuzzy modeling error and parameter estimation error. Then the assumptions made by previous literature that the optimal approximation error be square-integrable or the optimal approximation error be known are canceled. Also this approach has extended the application of supervisory control. By theoretical analysis, the closed-loop fuzzy control system is proven to be globally stable in the sense that all signals involved are bounded, with tracking error converging to zero.Secondly, using the Takagi-Sugeno(T-S) fuzzy systems as the approximator, the model reference decentralized adaptive control problem for a class of MIMO nonlinear systems with high-order interconnections is systematically studied. Considering two cases that the system has constant control gains or function control gams, two novel decentralized adaptive sliding mode control schemes are presented, respectively. Motivated by the principle of certainty equivalent control, adaptive estimation of unknown parameters is performed. And the neural networks are used to approximate the unknown functions on line, where the adaptive law of connection weights is embedded with fuzzy membership functions. Then the hybrid adaptive intelligent control is achieved through the natural integration of fuzzy logic and neural networks. At the same time, the decentralized control is guaranteed through some appropriate management of the interconnections. By Lyapunov method, the closed-loop control system is proven to be globally stable in the sense that all signals involved are bounded, with tracking error converging to zero.Thirdly, the adaptive fuzzy control problem for a class of SISO nonlinear systems with immeasurable state is studied. Two design schemes, i.e. direct fuzzy control and indirect fuzzy control, are proposed to improve the previous results and provide new approaches. In the direct fuzzy control case, a modulation function of observer-variables is adopted to keep the input-sets of fuzzy system always bounded. Then some propertiesabout the approximation capability of fuzzy logic systems are used, with canceling the assumption of boundedness of observer-variables. In the indirect fuzzy control case, the controller design is based on the backstepping techniques, so some restrictions of usual method are avoided. The stability of the closed-loop fuzzy control system is proven through constructing different Lyapunov functions.Lastly, the problem of tracking control for robot manipulators, where only the joint positions are measurable, is studied. A stable state observer is designed based on the principle of sliding mode control. Based on such observer and backstepping techniques, a new design scheme of robust tracking control is proposed. The damping term and modification parameters are adopted to restrain the observing error, then the converging speed of output tracking error is increased. Further more, the tracking error is proven to converge to zero through theoretical analysis.Through the research in th...
Keywords/Search Tags:fuzzy modeling, nonlinear systems, adaptive control, intelligent control, observer, stability
PDF Full Text Request
Related items