Robust neural network-driven control for multi-agent formation in the presence of Byzantine attacks and time delays

This paper presents an adaptive leader-follower formation control strategy for second-order nonlinear multi-agent systems with unknown dynamics. To handle system uncertainties, we used neural networks (NNs) to approximate and compensate for nonlinear effects. A key feature of our approach is its ability to deal with Byzantine attacks and time delays, which can disrupt coordination among agents. Unlike existing methods, our control strategy actively accounts for these challenges while ensuring stable formation tracking. Using Lyapunov stability theory, we proved that all system errors remain within a bounded range. Numerical simulations confirmed the effectiveness of our approach, showing that it successfully maintains formation control even in the presence of adversarial attacks and delays.