Distributed Fault-Tolerant Formation Control Design via High-Order Sliding Mode for a Team of Car-Like Vehicles

This study proposes a distributed fault-tolerant formation control scheme for a team of car-like vehicles with actuator faults. The proposed method includes a fault diagnostic scheme using a high-order sliding mode observer and a reconfigurable controller based on the super-twisting algorithm. The results showed that by the proposed controller, each vehicle can detect, diagnose, and accommodate actuator faults by itself. Two control strategies are proposed: 1) When a vehicle is exposed to a partial fault, the controller accommodates the fault and enables the vehicle to continue the mission with the rest of the team members in the formation. Furthermore, it allows healthy vehicles to re-adjust their motion and move in the way for compensating the reduced capability of the faulty vehicle with degraded performance in the formation mission; 2) In severe fault conditions, the proposed controller excludes the faulty vehicle and re-assigns its task to the rest of the healthy members to complete the mission. The re-assignment task for the team formation is formulated as an optimal assignment problem and is solved by the Simplex algorithm. The experimental tests have been conducted on the latest Quanser self-driving car (QCar) platform to verify the effectiveness of the proposed algorithm and ensure its capability to avoid accidents between team members.