Fault-Tolerant Control of Quadrotor UAV Formation Based on Global Graphical Game Theory*

A distributed optimal fault-tolerant controller for a formation system of quadrotor unmanned aerial vehicles is proposed via the global graphical game approach in this article. Firstly, the formation dynamic model and the fault model are simplified, and a fixed-time observer is designed to estimate actuator faults and external disturbances. Subsequently, the problem of fault-free formation control is reformulated as a global graphical game task, and the Hamilton-Jacobi-Bellman (HJB) equation is constructed to derive the local optimal controller. To address the solution of the HJB equation, an adaptive dynamic programming method is employed to approximate the neural network weights and obtain the global optimal control law. Moreover, by combining the local optimal controller and observer given above, a global optimal fault-tolerant controller is obtained, which achieves a global Nash equilibrium. Ultimately, a simulation serves to prove the applicability of the above theories.