Study On The Adaptive Control For A Class Of Semi-Strict Feedback Nonlinearly Parameterized Systems

In actual systems, there exists a large amount of uncertainties, such as parametric uncertainty and non-parametric uncertainty, in which the parametric uncertainty can be described by linear parametrization(LP) and nonlinear parametrization(NLP). There are many good results about adaptive control on the LP systems, however, adaptive control on the NLP systems is still a difficult and popular subject due to its high complexity and variability.This thesis deals with a class of semi-strict feedback nonlinearly parameterized sys-tems, and presents the relevant solutions. A dynamic surface controller (DSC) is presented to handle the case when the nonlinear uncertainties are bounded; besides, an adaptive ro-bust controller (ARC) based on immersion and invariance (I&I) is proposed for a stricter case. The results are as follows:(1)Towards a semi-strict NLP nonlinear system, an identification law based on I&I rather than certainty equivalence is proposed, which depends on some introduced tuning functions. By choosing the appropriate tuning functions, the law drives the estimation error between estimation and real value into the invariant manifold with zero identification error, so that the the estimation can converge.(2) Towards a class of nonlinearly parameterized systems with bounded nonlinear uncertainties, We design a dynamic surface controller combining with I&I adaptive laws. By adding a 1st order filter to the virtual controller, the DSC controller can prevent the complexity of the systems and consequently ensures the boundness of the status and output with a simple controller.(3) An adaptive robust controller based on I&I manifold method is proposed for a class of nonlinearly parameterized systems. Introduction of tuning functions in I&I into estimator for the unknown parameter in the system, gives the freedom in system controller design and ensures the asymptotic convergence of estimation error, the designed adaptive robust controller can not only deal with the nonlinearly parameterized uncertainties and exterior disturbances, but also guarantee the desired transient and final performance.