An Improved Power Steering System Based on Vector Control

Abstract

Electrically Power-Assisted Steering Systems (EPAS) have become the preferred choice selected by the majority of automotive Original Equipment Manufacturers (OEM). These systems use an electric motor to provide the required steering assist to the driver. Previously, these systems were using a DC motor to provide the required assistance. Permanent Magnet Synchronous Machines (PMSM) are replacing the traditional DC motors at the cost of increased system price and complexity. These systems often deploy high-resolution position sensors to accurately determine the rotor position at any given instance. These sensors often increase the total system price. Although Sensorless techniques can eliminate the use of a position sensor, they may not be convenient for applications requiring high dynamic response over the entire operating speed range such as the EPAS. This research aims to optimize the cost and complexity of the EPAS system when powered by a PMSM instead of a DC motor. This is achieved by replacing the high-resolution, expensive position sensor with a traditional set of low-resolution hall-effect sensors. A rotor position estimation algorithm is then used to obtain a high-resolution rotor position from the Hall effect sensors. Special focus was given to improve the estimator’s performance, especially during acceleration, deceleration, and direction reversal. For this reason, a complete model of the EPAS system was built using Matlab/Simulink which includes: the mechanical part, assist motor and the Electronic Control Unit (ECU) which controls the system. An experimental setup was then implemented around an actual EPAS unit. A customized ECU was built using an ARM-based Digital Signal Processor (STM32F407) and a special PCB which includes the Voltage Source Inverter (VSI) and other necessary electronics. A simulated mechanical load was built using a bike