When an additional task, such as avoiding obstacles, is required in the cooperative tracking control process, the leader's control signal is nonzero and time varying and accessible to any follower. On the other hand, the communicational and computational resources are limited for unknown nonlinear multiagent systems. This work develops a distributed robust event-triggered control algorithm to compensate for the noncooperation arising from the leader and reduce the resource consumption. Ultimately, our result is applied in multimanipulator systems with the time-varying leader, which implies that it can be further extended to practical multirobot applications with an additional task.

This article investigates the optimal cooperative tracking control problem for unknown nonlinear multiagent systems with a real dynamic leader and uncertainties. Its main difficulty lies in eliminating effect of the unmatched uncertainty and noncooperation in the event-triggered manner. To overcome this difficulty, a distributed robust event-triggered control algorithm is developed with the aid of a disturbance observer and robust approximate dynamic programming (ADP). Specifically, this problem is converted into a cooperative multiplayer game for the distributed auxiliary systems employing observer-based feedforward compensation, which contributes to reducing conservatism; via an adaptive event-triggered mechanism and a new weight tuning law, a distributed robust ADP approach is proposed with hope to obviate judging the existence of the saddle point beforehand. Moreover, the local neighbor tracking error and the weight estimation error are demonstrated to be uniformly ultimately bounded by means of Lyapunov stability theory. We also prove that Zeno behavior is excluded. A comparative simulation example for multiquadrotor systems is carried out to verify the effectiveness and superiority of our algorithm.