Observer-Based Optimal Position Control for Electrohydraulic Steer-by-Wire System Using Gray-Box System Identified Model

Abstract

Steer-by-wire (SBW) systems in a passenger car can improve vehicle steering capability and design flexibility by replacing the mechanical linkage between the steering wheel and front wheels by a control circuit. The steering controller, however, should provide good performance in response to driver's input signal. This includes fast response, absence of overshoot or oscillatory behavior, and good accuracy with minimal steady-state error. In this paper, an optimal control strategy based on observed system states is proposed and implemented on an electrohydraulic SBW system of a passenger car. First, a linear mathematical model is developed using gray-box system identification techniques. A standard input signal, pseudorandom binary sequence (PRBS), is designed to stimulate the system in the concerned bandwidth. Then, a linear-quadratic regulator (LQR) together with a full-state system observer is designed. Based on simulation, the LQR parameters and the observer poles are chosen to satisfy the aforementioned performance criteria for good steering. Finally, the control strategy is applied in a real-time environment to test the tracking capability, where the system is given high-rate reference signals (relative to the human rate of steering). The results show that the steering system tracks the reference signal with high accuracy even in the existence of high external force disturbances.