Adaptive Flocking of Multi-Agent Systems with Uncertain Nonlinear Dynamics and Unknown Disturbances Using Neural Networks*

Collective behavior of multi-agent systems brings some new problems in control theory and application. Especially, flocking problem of multi-agent systems with uncertain nonlinear dynamics and unknown external disturbances is a challenging problem. Some existing works assume that the intrinsic nonlinear dynamics of virtual leader is the same as those of the agents, which is unreasonable and impractical. To solve this issue, we consider an adaptive flocking problem of multi-agent systems with uncertain nonlinear dynamics and unknown external disturbances in this paper, where the intrinsic nonlinear dynamics of virtual leader is allowed to be different from the agents. Firstly, to approximate the uncertain nonlinear dynamics of each agent, an adaptive neural network is used, whose weights are updated online. Furthermore, an adaptive robust signal is designed to counteract the unknown external disturbances and neural network approximation errors, which is independent with the upper bound of the unknown external disturbances and neural network approximation errors. Moreover, an adaptive flocking control law is designed, which is proved that the flocking can be realized and the velocity errors converge to a small neighbor of the origin based on Lyapunov stability theory. Finally, the robustness and superiority of the proposed robust adaptive flocking control law are validated by two representative simulations.