Finding the Nearest Negative Imaginary System with Application to Near-Optimal Controller Design

TL;DR

This paper develops a method to find the closest negative imaginary (NI) system to a given non-NI system, enabling near-optimal controller design that guarantees robustness in feedback systems.

Contribution

It introduces a novel methodology to identify the nearest NI system to any LTI system, facilitating robust controller synthesis for NI systems.

Findings

Method to compute the nearest NI system from a non-NI system.

Application of the method to design near-optimal controllers.

Guarantees robustness in feedback control systems.

Abstract

The negative imaginary (NI) systems theory has attracted interests due to the robustness properties of feedback interconnected NI systems. However, a full output optimal controller-synthesis methodology, for such class of systems, is yet to exist. In order to develop a solution towards this problem, we first develop a methodology to find the nearest NI system to a non NI system. This later problem stated as follows: for any linear time invariant (LTI) system defined by the state space matrices $(A, B, C, D)$, find the nearest NI system, with the state space matrices $(A+\Delta_A,B+\Delta_B,C+\Delta_C,D+\Delta_D)$, such that the norm of $(\Delta_A,\Delta_B,\Delta_C,\Delta_D)$ is minimized. Then, this methodology will be used to find the nearest optimal controller for a given NI plant. In other words, for a given NI system, an optimal control methodology, such as LQG, is used to design an optimal controller that satisfy a particular performance measure. Then, the developed methodology of finding the nearest NI system is used, as a near-optimal control synthesis methodology, to find the nearest NI system to the designed optimal controller. Hence, the synthesized controller satisfy the NI property and therefore guarantee a robust feedback loop with the negative imaginary system under control.