Preliminary Results for LQR-Based Adaptive Learning Control

Document Type

Conference Proceeding

Publication Date

1-1-2025

Abstract

This paper proposes a novel low-complexity algorithm that integrates least-squares algorithm (LSA) and linear quadratic regulator (LQR) for adaptive learning control (ALC) to address the optimal control problem involving unknown system parameters. Compared with the existing results on reinforcement learning control (RLC) with high computational costs, the proposed algorithm has complexity in the order of n{2} with n being the order of the plant model, effectively removing bottlenecks inherent in the existing work on RLC- LQ R, which requires large number of training samples and multiple iterations. An important feature of the proposed algorithm is its capability of achieving parameter estimation and optimal control without requiring accurate system models. The obtained control gain ensures the feedback stability while being asymptotically convergent to the optimal solution. The simulation results validate the effectiveness of the proposed algorithm in reducing computational complexity, offering a novel approach for real time control of complex uncertain systems.

Publication Source (Journal or Book title)

7th International Conference on Industrial Artificial Intelligence Iai 2025

This document is currently not available here.

Share

COinS