Neural Network-Based Hierarchical Fault-Tolerant Affine Formation Control for Heterogeneous Nonlinear Multi-Agent Systems

Under a premise of universal rigidity, the affine formation control method based on stress matrix can solve the formation maneuvering problem well. However, a failure of the agent in the system can easily destroy this condition of rigidity. Consequently, a fault-tolerant affine formation control problem for multi-agent systems (MASs) with partial loss-of-effectiveness (PLOE) and bias faults is investigated. In this paper, a neural network-based hierarchical fault-tolerant affine formation (NN-HFAF) control strategy is proposed for heterogeneous nonlinear MASs. Firstly, some virtual systems are built as a link between leaders and followers. The virtual systems affinely locate their target positions in the formation maneuvers through the real-time positions of leaders. Then, an adaptive fault-tolerant control algorithm is designed for followers to tracking the virtual systems. It can effectively prevents the impact of a few agent failures from spreading to the whole network. And the system dynamics of agents in the network are considered to be heterogeneous. Moreover, radial basis function neural networks (RBF-NNs) are introduced to approximate the nonlinear functions of dynamic systems, the computational burden is reduced by adopting the single parameter learning mechanism. Finally, the numerical simulations are given to verify the efficiency of the proposed method.