Event-Triggered Multiple Dynamic Targets Formation Tracking Without Well-Informed Agent: A General Exploring Relationship
Abstract
In this article, the multiple dynamic targets formation tracking (MDTFT) problem is studied for multiagent systems. The objective is to drive locally connected agents to form the predefined time-varying formation while tracking the convex hull spanned by multiple targets. In existing results, agents are divided into well-informed and uninformed ones, where it is assumed that the well-informed agents must obtain the information of all the targets. However, in large-scale deployment scenarios, exploring all the targets is a daunting task for well-informed agents. To handle this problem, a new framework is designed to solve the MDTFT problems without a well-informed agent. Each agent is allowed to explore any number of targets depending on its own capability, namely, the general exploring relationship. Then, by using the adaptive mechanism and boundary layer technique, a fully distributed MDTFT algorithm with adaptive gains is constructed to avoid global information and control chattering. Further, a stochastic event-triggered MDTFT algorithm is specifically conceived, in which the continuous communication among agents can be avoided. Compared with the existing event-triggered schemes, the stochastic event-triggered scheme can significantly reduce the triggering times. Finally, numerical simulations are given to demonstrate the effectiveness of the proposed MDTFT algorithms.
