Optimal Formation Centroid Tracking Control of Multiagent Systems With Application to AGVs

This paper explores the time-varying optimal formation centroid tracking problem for leader-follower multiagent systems. The objective is to minimize the sum of local time-varying cost functions under the centroid tracking constraint, namely, the centroid of all followers is capable of tracking the leader's trajectory. A novel prescribed-time distributed extended state observer is proposed for the followers to estimate the leader's trajectory and control input. Furthermore, to tackle the issue that the local cost function is only known individually, the prescribed-time dynamic average consensus approach is employed to obtain the optimal time-varying Lagrange multiplier. The optimal control protocol is proposed by utilizing the gradient of the Lagrange function as feedback error, and the adaptive gain method is incorporated to ensure that each follower can track the optimal time-varying trajectory. Eventually, the efficacy of the proposed method is confirmed through the numerical simulations and the practical experiment with multiple unmanned ground vehicles (UGVs).