Development of Supervisory and Adaptive PID Control for Nonlinear MIMO Systems(I)

Project: National Science and Technology CouncilNational Science and Technology Council Academic Grants

Project Details

Abstract

This project aims to propose a fuzzy-model based two-level robust adaptive control strategy for a class of multi-input multi-output (MIMO) nonlinear systems. Traditional adaptive control methods usually rely on simplified mathematical models. These methods often lack the guarantee of validity and magnitude-constrains of adaptation gains, render adaptation control rarely found in practical systems. In recent years, we have successfully developed a two-level adaptive control scheme for single-input single-output nonlinear systems. We plan to extend the results to MIMO systems and experimentally verify the strategy in this project. We will combine evolutionary computation and steepest decent methods to identify the plants, and use the projection adaptation law to ensure the convergence of gains, in order to take stability, modeling error, transient performance, and external disturbances into account. The first year of the 3-year project is devoted to develop an evolutionary identification strategy and a real-time identification method for the MIMO nonlinear plants. The identified models will be in the form of fuzzy models. In the second year, we will extend the research results of identification to the design of two-level adaptive control law for MIMO nonlinear systems. First of all, a supervisory controller will be derived based on the identified model and Lyapunov function. The controller is activated according to a switching index to enhance system stability and improve transient performance. Within the inner level, a PID adaptive controller will be designed using the projective method, together with a feed-forward controller, both taking time-delay into consideration. The results will be implemented on a tracking control system in the third year, specifically, a three DOF robot arm. The system is with significant nonlinearity, and strong coupling between state variables. The research will be accomplished with practical modifications and performance comparisons with other state-of-the-art control techniques.

Project IDs

Project ID:PB9709-4333
External Project ID:NSC97-2221-E182-025
StatusFinished
Effective start/end date01/08/0831/07/09

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.