Command Governors with Inexact Optimization and without Invariance

TL;DR

This paper introduces a modified command governor that operates effectively with inexact optimization and without invariant sets, broadening its practical applicability in constrained control systems.

Contribution

A simple modification to command governors that allows operation with inexact optimization and without invariant sets, enhancing their practical use.

Findings

Enables command governors to work with inexact optimization.

Removes the need for invariant constraint sets.

Demonstrates effectiveness through numerical examples.

Abstract

Reference and command governors are add-on schemes that augment nominal closed-loop systems with the capability to enforce state and control constraints. They do this by monitoring and modifying, when necessary, the reference command. Existing command governors do this by solving at each sampling time a quadratic programming problem to find a modified reference closest to the original command such that the current state and the modified reference pair are constraint admissible. In this paper, we show that a simple modification of the basic command governor enables it to operate with inexact optimization and even without requiring invariance of the constraint admissible set. Thus this modification significantly extends the applicability of the reference and command governors to practical problems where finding invariant sets may be problematic and where exact optimization may not be feasible due to reliability of the optimizers or limited computing power. Numerical examples are reported which illustrate the approach.