Geometric Region-Based Swarm Robotics Path Planning in an Unknown Occluded Environment

摘要

This article presents a geometrical region-based shape control methodology for navigating a cohesive swarm-robotic structure toward the goal even in a field occluded by unknown obstacles. In this control approach, initially, the robotic swarm is conceived to lie within a well-defined virtual circular region thus preserving a strict interagent cohesiveness among them. However, during the progression, for evading severely constricted obstacles, the virtual circle has been allowed to change its shape and in the process, varied elliptical shapes are made to evolve. In essence, for a collision-free solution, this shrinking aspect (from circle to ellipse) depends entirely on the number of agents in the swarm and at the same time also reliance on the sensed distance between two nearest obstacles through which the shrunken circle or the virtual ellipse will be able to pass. Consequently, shape switching is a dynamic as well as a stochastic process throughout the journey of the swarm. For achieving these objectives, a two-level hierarchical control strategy has been employed. Moreover, during aggregating toward the target, the actuation failure of any agent or agents may occur. In this perspective, the proposed control law has been updated adaptively throughout the route such that agent failure does not encumber the mission. Finally, the extensive simulation results along with the hardware experimentation are provided to demonstrate the efficacy of the proposed scheme.