Distributed Prescribe Performance Control Based on Adaptive Neural Network Strategy for Multi-AAVs Under Cyber-Attacks

This article proposes a distributed finite-time prescribed performance (FTPP) control strategy with the adaptive neural network (NN) compound approximation method to achieve the cooperative control of multiple autonomous aerial vehicle systems (MAAVSs). The deceptive and injective cyber-attacks on sensors and actuators of MAAVS are considered. The attitude controller is designed based on the distributed backstepping law, by which the stepwise virtual control signals were designed to control the attitude states of the AAVs. The FTPP control scheme predefines the error boundary and the convergence time to constrain the cooperative tracking errors in a small neighborhood and the settling time. Then, the MAAVS are controlled to track the given reference trajectory rapidly. In addition, to obtain information on the actuator and sensor attacks, the NN method is synthesized with the adaptive law to approximate the attacks and alleviate the undesirable effects. The Lyapunov method was employed to analyze and prove the stability of distributed cooperative errors under AAV communication network described by graph theory. The conducted experiments verify the effectiveness and the superiority of the proposed control algorithm.