Dynamics and Model-Based Control of Electric Power Steering Systems

Abstract

Many automobile manufacturers are switching to Electric Power Steering (EPS) systems for their better performance and cost advantages over traditional Hydraulic Power Steering (HPS) systems. EPS compared to HPS offer lower energy consumption, lower total weight, and package flexibility at no cost penalty. Furthermore, since EPS systems can provide assistance to drivers independent of the vehicle driving conditions, new technologies can be implemented to improve the steering feel and safety, simultaneously.

In this thesis, a neuromusculoskeletal driver and a high-fidelity vehicle model are developed in MapleSim to provide realistic simulations to study the driver-vehicle interactions and EPS systems. The vehicle model consists of MacPherson and multilink suspensions at front and rear equipped with a column-type EPS system. The driver model is a fully neuromusculoskeletal model of a driver arm holding the steering wheel, controlled by the driver's central nervous system. A hierarchical approach is used to capture the complexity of the neuromuscular dynamics and the central nervous system in the coordination of the driver's upper extremity activities. The proposed motor control framework has three layers: the first layer, or the path-planning layer, plans a desired vehicle trajectory and the required steering angles to perform the desired trajectory, the second layer (or the force distribution controller) actuates the musculoskeletal arm, and the final layer is added to ensure the precision control and disturbance rejection of the motor control units.

The overall goal of this thesis is to study vehicle-driver interactions and to design a model-based EPS controller that considers the driver's characteristics. To design such an EPS controller, the high-fidelity driver-vehicle model is simplified to reduce the computational burden associated with the multibody and biomechanical systems. Then, four driver types are introduced based on the physical characteristics of drivers such as age and gender, and the corresponding parameters are incorporated in the model. Last but not least, a new model-based EPS controller is developed to provide appropriate assistance to each of the predefined driver types. To do this, the characteristic curves are tuned using a systematic optimization procedure to provide appropriate assistance to drivers with different physical strength, in order to have a similar road and steering feel. In this thesis, it is recommended that muscle fatigue be used as a measure of steering feel. Then, based on the tuned EPS characteristic curves, an observer-based optimal disturbance rejection controller, consisting of a linear quadratic regulator controller and a Kalman filter observer augmented with a shaping filter, is developed to deliver the assistance while attenuating external disturbances. The results show that it is possible to develop a model-based EPS controller that is optimized for a given driver population.