Neural-Network-Based Event-Triggered Formation Tracking for Nonlinear Multi-UAV Systems With Switching Topologies Under DoS Attacks

A time-varying formation tracking (TVFT) control method under denial-of-service (DoS) attacks is proposed for a class of multi-uncrewed aerial vehicle (UAV) system with unknown nonlinearity. The neural network approximation is utilized to mitigate the impact of nonlinear dynamics on the tracking performance. Meanwhile, the switching topologies are adopted to ensure the reliability of the communication topology. Considering the resource constraints of the system communication network, an improved integral event-triggered mechanism is devised to further decrease the triggering frequency in comparison to the traditional event-triggered one. By using the multi-Lyapunov function, the tracking error is proved to be bounded. Simulations convincingly demonstrate the effectiveness and merits of the developed TVFT and the integral event-triggered method. Note to Practitioners—Multi-UAV systems are widely utilized in both civil and military domains, with communication networks serving as crucial channels for information exchange among UAVs. However, these communication networks are often vulnerable to various types of attacks. Therefore, the control problems of multi-UAV systems under DoS attacks are of great importance. A switching event-triggered control protocol is developed for multi-UAV systems to facilitate effective formation tracking in the presence of DoS attacks. This protocol employs switching topologies to enhance the reliability and flexibility of the communication network. The improved event-triggered mechanism effectively balances the system performance and the communication consumption.