Active Suspension Control System Using H-Infinity Mixed Sensitivity Technique With Automatic Weights Selection

Abstract

Weight selection is critical in the determination of the sensitivity function (which defines performance in terms of disturbance rejection) and the complementary sensitivity function (which specifies the set-point tracking of the system). The selection of weight in conventional H-infinity control is rigorous and requires high designer’s experience. This paper presents the development of a low cost and efficient robust controller for the active suspension system using H-infinity (H_∞) mixed sensitivity with automatic weights selection.  In this work, a higher level designer’s specification for the active suspension of a half car is mapped into a multi-objective optimization problem; the Multi-Objective Particle Swarm Optimization (MOPSO) algorithm is deployed to solve the well-posed control optimization problem in order to achieve the stated design specification. The formulation of the control problem involves the use of weighing functions with tunable parameters. Mixed sensitivity technique is used to calculate the controller gain such that it stabilizes the system and minimizes the infinite norm of the system. Analyses of the robust stability and robust performance were carried out after simulation and the results indicate a good robust stability and performance. For the ride comfort, the developed controller outperforms conventional H-infinity using Algebraic Riccati Equation (ARE) and H₂ controller using LMI in terms of the rise time and settling time. The computation time (a reflection of cost of design) is reduced by reduction of order of the controller which justifies the ease of its implementation.