Chance-Constrained Controller State and Reference Governor

TL;DR

This paper extends the controller state and reference governor (CSRG) framework to stochastic systems with chance constraints, providing algorithms and theoretical guarantees for constraint enforcement, stability, and convergence.

Contribution

It introduces a stochastic CSRG algorithm that enforces chance constraints and analyzes its stability and convergence properties.

Findings

CSRG achieves larger constrained domains than traditional schemes.

Theoretical guarantees for chance-constraint satisfaction and stability.

Application demonstrated in aircraft flight control.

Abstract

The controller state and reference governor (CSRG) is an add-on scheme for nominal closed-loop systems with dynamic controllers which supervises the controller internal state and the reference input to the closed-loop system to enforce pointwise-in-time constraints. By admitting both controller state and reference modifications, the CSRG can achieve an enlarged constrained domain of attraction compared to conventional reference governor schemes where only reference modification is permitted. This paper studies the CSRG for systems subject to stochastic disturbances and chance constraints. We describe the CSRG algorithm in such a stochastic setting and analyze its theoretical properties, including chance-constraint enforcement, finite-time reference convergence, and closed-loop stability. We also present examples illustrating the application of CSRG to constrained aircraft flight control.