Control Design under Actuator Saturation and Multi-Rate Sampling

TL;DR

This paper develops a hybrid control design approach for stabilizing systems with actuator saturation and asynchronous multi-rate measurements, providing explicit stability conditions and a method to maximize the region of attraction.

Contribution

It introduces a novel hybrid control framework with matrix inequality conditions for stability under saturation and multi-rate sampling, including a design procedure to enlarge the basin of attraction.

Findings

Proposed sufficient conditions for regional exponential stability.

Explicit estimates of the basin of attraction as ellipsoidal sets.

A semidefinite programming based design method to maximize the stability region.

Abstract

The problem of designing a stabilizing feedback controller in the presence of saturating actuators and multi-rate (asynchronous) aperiodic state measurements is studied. Specifically, we consider a scenario in which measurements of the plant states are collected at the controller end in a sporadic and asynchronous fashion. A hybrid controller is used to perform a fusion of measurements sampled at different times. In between sampling events, the controller behaves as a copy of the plant and provides a feedback control signal based on the reconstruction of the plant state. The presence of saturation at the plant input limits the value of the components of this signal to a bounded range. When a new measurement is available, the controller state undergoes an instantaneous jump. The resulting system is augmented with a set of timers triggering the arrival of new measurements and analyzed in a hybrid systems framework. Relying on Lyapunov tools for hybrid systems and techniques for control design under saturation, we propose sufficient conditions in the form of matrix inequalities to ensure regional exponential stability of a closed-set containing the origin of the plant, i.e., exponential stability with a guaranteed region of attraction. Specifically, explicit estimates of the basin of attraction are provided in the form of ellipsoidal sets. Leveraging those conditions, a design procedure based on semidefinite programming is proposed to design a stabilizing controller with maximized size of the basin attraction. The effectiveness of the proposed methodology is shown in an example.