H∞ Cluster Formation Control of Networked Multiagent Systems With Stochastic Sampling

This article is concerned with the
$H_{\infty }$
cluster formation control of a multiagent system (MAS) with stochastic sampling in network environments. First, based on a directed communication topology with acyclic partition, agents are separated into several clusters. The agents moving in the same cluster are expected to achieve a desired formation collaboratively, while the agents in different clusters have different formation patterns. Second, the external disturbance for each agent is taken into account. The associated
$H_{\infty }$
cluster formation control, whose performance bound can be obtained by considering both formation information and cluster characteristics, is introduced to measure the disturbance attenuation ability. Third, by casting a stochastic sampled-data-based
$H_{\infty }$
cluster formation problem into an
$H_{\infty }$
control problem of a stochastic system, an
$H_{\infty }$
cluster formation criterion is derived for the MAS with external disturbance. Finally, a lower-dimensional cluster formation criterion is obtained for the disturbance-free MAS. Cluster formation performance analysis testifies the effectiveness of the proposed design methods.