Output Feedback Control Based on State and Disturbance Estimation

TL;DR

This paper introduces a novel output feedback control method that estimates and compensates for disturbances in multi-input multi-output systems without requiring specific disturbance assumptions, enhancing robustness and stability.

Contribution

It extends disturbance estimation to general MIMO systems without relying on matching conditions or canonical forms, and provides a stability framework for the control design.

Findings

Effective disturbance rejection demonstrated on a first-order system.

Successful stabilization of an inverted pendulum under uncertainties.

The control method does not require disturbance independence or matching conditions.

Abstract

Recently developed control methods with strong disturbance rejection capabilities provide a useful option for control design. The key lies in a general concept of disturbance and effective ways to estimate and compensate the disturbance. This work extends the concept of disturbance as the mismatch between a system model and the true dynamics, and estimates and compensates the disturbance for multi-input multi-output linear/nonlinear systems described in a general form. The results presented do not need to assume the disturbance to be independent of the control inputs or satisfy a certain matching condition, and do not require the system to be expressible in an integral canonical form as required by algorithms previously described in literature. The estimator and controller are designed under a state tracking framework, and sufficient conditions for the closed-loop stability are presented. The performance of the resulting controller relies on a co-design of the system model, the state and disturbance observer, and the controller. Numerical experiments on a first-order system and an inverted pendulum under uncertainties are used to illustrate the control design method and demonstrate its efficacy.