Algorithms and Experiments on Flocking of Multiagents in a Bounded Space

Abstract

In most scenarios encountered in both natural systems and engineering applications, collective motions are always confined in a bounded space. The conventional flocking formulation with velocity consensus does not apply in these scenarios as agents with a constant velocity that unavoidably move out of any bounded space. In this brief, the conventional flocking formulation and algorithms are generalized to address the issues caused by bouncing boundaries of a bounded space. In addition, the theoretical design was successfully implemented in a group of real robots such that it experimentally achieved a desired flock in a bounded space.