Optimal Fault Estimation and Fault-tolerant Control for Spacecraft Formation via Stackelberg Differential Game approach
Jing Huang, Yizhen Meng, Dongfang Zhu, Jun Sun · 2023
In this work, a deep investigation has been conducted into the optimal fault estimation and accommodation problem for spacecraft formation flying, which has resulted in the formulation of a fault-tolerant formation control scheme using the framework of Stackelberg differential game theory. Within the framework of the Stackelberg game, the role of the leader is assumed by the fault observer, while the leader is impersonated by the spacecraft formation system, thereby establishing sequential decisions between the follower and the leader. To achieve the Stackelberg equilibrium, an auxiliary controller is designed for the follower, enabling a non-cooperative game with the control strategy of the leader. To facilitate the online updating of the schemes of the follower and the leader, the reinforcement learning control scheme is applied, utilizing two Actor-Cntic neural networks for the observer and spacecraft formation system, respectively. Ultimately, the simulation is carried out to achieve the optimal fault estimation and accommodation targets, thereby illustrating the efficiency and applicability of the proposed approach.