Haptic Feedback Control Methods for Steering Systems

Licentiate thesis, 2019

Haptic feedback from the steering wheel is an important cue that defines the steering feel in the driver-vehicle interaction. The steering feedback response in an electric power assisted steering is primarily dependent on its control strategy. The conventional approach is open loop control, where different functions are implemented in a parallel structure. The main drawbacks are: (a) limited compensation of the hardware impedance, (b) hardware system dependent steering feedback response and (c) limitation on vehicle motion control request overlay. This thesis investigates closed-loop control, in which the desired steering feedback response can be separated from the hardware dynamics. Subsequently, the requirements can be defined at the design stage.

The closed-loop architecture constitutes of a higher and lower level controller. The higher level control defines the reference steering feedback, which should account for both driver and road excitation sources. This thesis focuses on the driver excitation, where a methodology is proposed for developing such a reference model using the standard vehicle handling maneuvers. The lower level control ensures: (a) reference tracking of the higher level control, (b) hardware impedance compensation and (c) robustness to unmodeled dynamics. These interdependent objectives are realized for a passive interaction port driving admittance. The two closed-loop possibilities, impedance (or torque) and admittance (or position) control, are compared objectively. The analysis is further extended to a steer-by-wire force-feedback system; such that the lower level control is designed with a similar criteria, keeping the same higher level control.

The admittance control is found limited in performance for both the steering systems. This is explained by a higher equivalent mechanical inertia caused by the servo motor and its transmission ratio in electric power assisted steering; and for steer-by-wire force-feedback, due to the uncertainty in drivers' arm inertia. Moreover, it inherently suffers from the conflicting objectives of tracking, impedance compensation and robustness. These are further affected by the filtering required in the admittance lower level control. In impedance control, a better performance is exhibited by its lower level control. However, the required filtering and estimation in the impedance higher level control is its biggest disadvantage. In closed-loop setting, the angular position overlay with a vehicle motion control request is also relatively easier to realize than open loop.