Active Fault-Tolerant Control of a Four-Wheel Independent Steering System Based on the Multi-Agent Approach

Abstract

In this paper, we present an active fault-tolerant, vehicle motion control, based on the multi-agent approach for a four-wheel independent steering system. This improves handling stability and active safety of the vehicle in the case of actuator faults. Firstly, an eight-degree-of-freedom four-input vehicle model is established by using the vector transformation method, which takes the four-wheel angle as the input and the yaw rate and the sideslip angle, obtained by the lateral force of a single wheel, as the outputs. Then, we established the topological structure of the four-wheel independent steering subsystems based on graph theory. Aiming at the known fault conditions, an adaptive sliding mode active fault-tolerant control based on the multi-agent four-wheel independent steering system is proposed, in which the unmodeled part of the system is estimated using adaptive methods. For unknown fault conditions, an active fault-tolerant control of the four-wheel independent steering system based on state observer is proposed. An extended state observer with sliding mode surface function as a state variable is designed to estimate actuator faults. Finally, it was demonstrated that the designed state observer and distributed active fault-tolerant control method reduced the impact of actuator failures on vehicle steering and improved the stability of the steering process based on the Lyapunov theory. The hardware in the loop experimental results validate the effectiveness of the proposed novel method.