Loading [MathJax]/extensions/MathZoom.js
Regulation-Triggered Batch Least-Squares Identifier based Adaptive Control with Unknown Control Coefficients | IEEE Conference Publication | IEEE Xplore

Regulation-Triggered Batch Least-Squares Identifier based Adaptive Control with Unknown Control Coefficients


Abstract:

This paper focuses on regulation-triggered batch least-squares identifier based adaptive control for nonlinear systems with unknown control coefficients. Batch least-squa...Show More

Abstract:

This paper focuses on regulation-triggered batch least-squares identifier based adaptive control for nonlinear systems with unknown control coefficients. Batch least-squares identifiers (BaLSI) are designed to estimate the unknown parameters, while a regulation-triggered condition is proposed to activate the BaLSI. Previous appeared excitation are stored and utilized for parameter updates at the triggering moments, based on which the persistently exciting (PE) conditions are no longer needed. Effects of unknown parameters can be eliminated after finite times of parameter updates. Compared with the existing results, the unknown control coefficients are considered, which may lead to singular control input. It implies that as the estimate of control coefficient approaches zero, the control input tends towards infinity. To address the aforementioned issue, two distinct control schemes have been proposed, which pertain to the parameter update law and the control law respectively.
Date of Conference: 25-27 May 2024
Date Added to IEEE Xplore: 17 July 2024
ISBN Information:

ISSN Information:

Conference Location: Xi'an, China

Funding Agency:


I. Introduction

Adaptive control is an important branch of control field, which has been developed for several decades. Huge amount of valuable adaptive control schemes have been proposed, which are detailed introduced in [1]–[6] and the references therein. These adaptive control schemes are classified into "Lyapunov-based" and "estimation-based" (as known as "modular-based") approaches in [1]. The "Lyapunov-based" adaptive control schemes are designed to guarantee some desired properties of the Lyapunov functions, such as with α of a class K function. Although desired performance of the closed-loop systems and boundedness of the estimates are both guaranteed in theory, the effects of rapidly varying estimation and "differential explosion" problem of high-order systems are two disadvantages. These drawbacks bring difficulty to the control design in practical applications. "Estimation-based" adaptive control includes least-squares and gradient approaches, which are more effective in practical applications. However, the proof and analysis of "estimation-based" schemes are quite challenging, and it always depends on the persistently exciting (PE) conditions.

References

References is not available for this document.