Based Data Model-Free Adaptive Control For A Class Of MIMO Nonlinear Systems

In recent years, human society has entered the age of big data. With the rapid development of information sciences and technology, production technology, production equipments and industrial processes are also becoming more and more complex. Therefore, the traditional mathematical model methods for analyzing and controlling the complex systems, which are used to establish the physical and chemical theories, have become infeasible. Due to modern digital sensors, digital storage, digital communication technology to universal use, the dynamical systems generate large amounts of data every day, which reflect various information. Therefore, effectively use these online and offline data to combine data mining, pattern recognition and computer control technology has become another challenging problem in control theory.This thesis is used data-based to develop model-free adaptive control for a class of MIMO nonlinear systems, and considered the external disturbance existing in output. The main contributions of this thesis are listed as follows:First, in the process of learning the control theory and system engineering, the study on system properties is an important subject. This thesis first develop the based-data method, which only use measurement input and output data, to analysis the state controllability, output controllability and state observability. Compared with the traditional model- based methods, data- based methods have the advantages of higher calculation precision and lower computational complexity.Second, the model-free adaptive control method is designed mainly for the nonlinear discrete time system. The main characteristic only needs input and output data to design the controller of the system. This thesis develops the MIMO nonlinear discrete-time system and establishes three forms of linear dynamic models, namely Compatible Form the Dynamic Linearization, Partial Form the Dynamic Linearization and Full Form the Dynamic Linearization. Based three forms of linear dynamic models and considered the error of the measurement output data. This thesis respectively designs the corresponding controllers and analyzes robust convergence and stability.Finally, this thesis present theoretical analysis and computer simulation to demonstrate free model adaptive control method is feasible.