Jose Velazquez Alcantar, Donald Margolis, and Olugbenga Anubi
A detailed nonlinear bond graph model of a column mounted electric power steering system connected to a bicycle car model is developed to analyze the dynamics of the electromechanical system in order to improve steering feel via automatic control. A reduced order linear model is then derived from the nonlinear model for linear control development. First, a simple reference model is developed in order to create an ideal Steering Torque vs Steering Angle hysteresis curve whose shape can be modified by the user. An ad-hoc linear controller is then developed in order to track the ideal steering torque signal created by the model reference. It is shown that the controller tracks the reference model’s torque signal while remaining under the power restrictions of the physical system. Furthermore, because the model reference can be tuned according to the demand of the user, it is also shown that this control strategy can help improve steering feel by tracking any Steering Torque vs Steering Angle hysteresis curve desired. Finally, the hysteresis tracking controller is benchmarked against a simplified generic standard controller commonly used in most vehicles. It is shown that the tracking controller provides the same performance, and in cases better performance, as the standard controller with the added benefit of only requiring four tuning parameters as opposed to the standard controller’s seventeen tuning parameters.