Multi-Objective Model Predictive Control for Vehicle Active Suspension System

Abstract

This paper puts forward a multi-objective model predictive control scheme in order to address the conflicting control objectives of a vehicle active suspension system. The key problem in designing an active suspension controller is that the controller has to realize a feasible control input that can satisfy the competing control requirements such as ride comfort, suspension travel and road handling. Hence, in this work, these constraints are integrated into an optimal control framework and a finite horizon model predictive controller is used to solve the multi-objective cost function. The key advantage of the proposed scheme is that the model predictive control design finds the optimal control input by solving the discrete time algebraic Riccati equation. This guarantees not only a robust closed loop system but also a realizable control effort, without violating the hard constraints of the active suspension system. The proposed model predictive control design is experimentally validated on a laboratory scale quarter car suspension system using hardware-in-loop testing. The performance of the model predictive control scheme is compared with the one of the unconstrained linear quadratic regulator and tested for four realistic road profiles. The experimental results substantiate that the suspension system controlled by the model predictive controller offers better ride comfort and road handling features when compared to the conventional linear quadratic regulator.