Leveraging area bounds information for autonomous decentralized multi-robot exploration

Abstract

h i g h l i g h t s• A uniform, decentralized, potential-field based approach to dispersing a team of robots to explore an area quickly.• Leverages knowledge of the overall bounds of the area to be explored and includes a monotonic coverage factor in the field equations to avoid minima.• Allows robot team members to become disconnected from and reconnected with the team and ongoing dispersion strategy.• Introduces performance metrics for team speedup and efficiency in the case of adding members to teams.• Presents both simulation and experimental robot results. a r t i c l e i n f o t r a c tThis paper proposes a simple and uniform, decentralized approach to the problem of dispersing a team of robots to explore an area quickly. The Decentralized Space-Based Potential Field (D-SBPF) algorithm is a potential field approach that leverages knowledge of the overall bounds of the area to be explored. It includes a monotonic coverage factor in the potential field to avoid minima, realistic sensor bounds, and a distributed map exchange protocol.The D-SBPF approach yields a simple potential field control strategy for all robots but nonetheless has good dispersion and overlap performance in exploring areas with convex geometry while avoiding potential minima. Both simulation and robot experimental results are included as evidence, and performance, speedup and efficiency metrics for each are presented.

Thorough exploration of complex environments with a space-based potential field

“…The approach taken was based on the Space-Based Potential Field (SBPF) method of Liu & Lyons [3]. That approach fails to thoroughly explore corners and niches because of the repulsive effect of walls and obstacles.…”

“…To demonstrate the ESBPF algorithm on physical hardware, a map with a single partitioned space, similar to that used in Liu & Lyons [3] was used. The room consisted of a 3.5 by 4 meter space, with a small "wall" obstacle, about one meter long, dividing the space into two "rooms" connected by a larger corridor.…”

“…The benefits include reducing the danger to first response personnel as well as allowing disaster victims to be treated more quickly. The Space-Based Potential Field (SBPF) algorithm developed by Liu & Lyons [3] is a decentralized control strategy which uses the same straightforward control scheme for each team member, but nonetheless leverages each team member effectively. It allows team members, even sub-teams, to become disconnected and reconnected during exploration as needed.…”

Distributed Multirobot Exploration Based on Scene Partitioning and Frontier Selection

“…Liu and Lyons [14] explain a multirobot exploration method for unknown scenarios. The method is based on creating a potential field that guides robots to unexplored areas and away from explored areas or other robots.…”

A Cognitive Agent-based Model for Multi-Robot Coverage at a City Scale

“…Similarly, in Zheng et al (2010), the authors have proposed Space-Based Potential Field technique, which focus on dispersion of robots in the environment, while avoiding overlap performance and the local minima. Recently decentralized variation of this technique (Liu and Lyons 2015) is proposed, which includes monotonic factor for coverage to avoid local minima problem.…”