Rear-wheel steering – state of the art control creation

Nowadays, the integration of various electronic systems in the chassis of vehicles for active safety purposes is widespread and ever increasing in the automotive industry. Motivated by this trend, this master thesis aims at developing an active rear-wheel steering (RWS) controller implemented on a virtual vehicle based on the know-how of Danisi Engineering company, headquartered in Nichelino (TO). For the controller’s creation, various classical and optimal control methods are explored and realized. An initial review displays the growing interests from major car manufacturers and researchers since the eighties regarding RWS systems’ realization and their control design. Benefits in terms of maneuverability and stability are obtained from proposals of RWS control based on different theories for various typologies of RWS systems. The vehicle of subject is modelled as a single-track model with nonlinear axle lateral force characteristics, essentially capturing the vehicle’s lateral dynamic behaviors. Starting from setting the desired control goals in terms of desired yaw rate and understeering characteristics, multiple controllers based on different control theories are created in MATLAB/Simulink, among them PI, LQI and MPC control. They all improve the vehicle’s lateral dynamic responses by exerting an active RWS angle from an input of yaw rate tracking error. The effectiveness of proposed control structures is verified from simulations of several typical steady-state and transient maneuvers in MATLAB/Simulink: ramp steer, step steer and sine sweep steer are considered. The controllers are subsequently tuned to have appropriate parameters able to achieve desired performance goals. Then, comparisons between the passive and controlled vehicle’s responses are made for every maneuver with each controller configuration. Simulation results show that the proposed RWS controllers effectively enhance the vehicle’s cornering responses and achieve their design goals.