ANALYTICAL REDUNDANCY BASED FAULT TOLERANT CONTROL OF A STEER-BY-WIRE SYSTEM

摘要

This paper presents a novel observer-based analytical redundancy for a steer-by-wire (SBW) system. In order to achieve high level of reliability for a By-Wire system, double, triple, or even quadruple redundant sensors, actuators, communication networks, and controllers are needed. But this added hardware increases the overall cost of the vehicle. This paper utilizes a novel analytical redundancy methodology to reduce the total number of redundant road-wheel angle (RWA) sensors in a triply redundant RWA-based SBW system, while maintaining a high level of reliability. The self-aligning torque at road-tire interface due to the steering dynamics has been modeled as a function of the linear vehicle states. A full state observer was designed using the combined model of the vehicle and SBW system to estimate the vehicle body side slip angle. The steering angle was then estimated from the observed and measured states of the vehicle (body side slip angle and yaw rate) as well as the current input to the SBW electric motor(s). With at least two physical road-wheel angle sensors and the analytical estimation of the RWA value (which replaces the third physical sensor), a fault detection and isolation (FDI) algorithm was developed using a majority voting scheme. The FDI algorithm was then used to detect faulty sensor(s) in order to maintain safe drivability. The proposed analytical redundancy based fault detection & isolation algorithms and the linearized vehicle model were modeled in SIMULINK. Simulation of the proposed algorithm was performed for both single and multiple sensor faults. Simulation results show that the proposed analytical redundancy based fault detection and isolation algorithm provides the same level of fault tolerance as in an SBW system with full hardware redundancy against single point failures.