Prescribed-time containment of nonlinear multiagent systems via learning control approach

This paper addresses the learning-based practical prescribed-time containment control problem of multiagent systems with uncertain nonlinearity. To handle different reference trajectories generated by multiple leaders, a practical predefined-time distributed observer is designed for each follower to estimate the ideal reference trajectory. Moreover, a time-varying constraint function is constructed to constrain the tracking error between the follower output and the estimated trajectory. Then, based on the constrained tracking error, a learning controller is designed under the backstepping framework, where neural networks are utilised to directly learn ideal control inputs. Subsequently, the practical prescribed-time stability of the proposed scheme is verified by Lyapunov stability analysis. Finally, the effectiveness of the control algorithm is validated through the Nomoto ship model.