Event-Triggered Optimal Cooperative NN Control for UAV Systems with Deception Attacks

Abstract

In this paper, an event-triggered optimal cooperative control scheme is proposed for six-rotor unmanned aerial vehicles subject to deception attacks by using the reinforcement learning algorithm. Based on the identifier-actor-critic frame, the identifier, actor and critic neural networks are employed to estimate unknown system dynamics, realize control actions and evaluate the system performances, respectively. In addition, an event-triggered strategy is introduced to reduce the waste of communication resources, which also can exclude the Zeno behavior. Then, a compensation control technique based on the disturbance observer is introduced to compensate for the deception attacks and the external disturbances. The proposed optimal cooperative control strategy guarantees that all signals of the six-rotor unmanned aerial vehicles are semi-globally uniformly ultimately bounded, and the outputs of the six-rotor unmanned aerial vehicle systems can achieve consensus with the desired accuracy. Finally, the simulation results are presented to test the effectiveness of the proposed control methods.