Flocking of multi-agents in constrained environments

Abstract

Flocking, arguably one of the most fascinating concepts in nature, has in recent times established a growing stature within the field of robotics. In this paper, we control the collective motion of a flock of nonholonomic car-like vehicles in a constrained environment. A continuous centralized motion planner is proposed for split/rejoin maneuvers of the flock via the Lyapunov-based control scheme to anchor avoidance of obstacles intersecting the paths of flockmates. The control scheme inherently utilizes the artificial potential fields, within a new leader-follower framework, to accomplish the desired formations and reformations of the flock. The effectiveness of the proposed control laws are demonstrated through computer simulations.