Efficient Boustrophedon Multi-Robot Coverage: an algorithmic approach

Abstract

This paper presents algorithmic solutions for the complete coverage path planning problem using a team of mobile robots. Multiple robots decrease the time to complete the coverage, but maximal efficiency is only achieved if the number of regions covered multiple times is minimized. A set of multi-robot coverage algorithms is presented that minimize repeat coverage. The algorithms use the same planar cellbased decomposition as the Boustrophedon single robot coverage algorithm, but provide extensions to handle how robots cover a single cell, and how robots are allocated among cells. Specifically, for the coverage task our choice of multi-robot policy strongly depends on the type of communication that exists between the robots. When the robots operate under the line-of-sight communication restriction, keeping them as a team helps to minimize repeat coverage. When communication between the robots is available without any restrictions, the robots are initially distributed through space, and each one is allocated a virtually-bounded area to cover. A greedy auction mechanism is used for task/cell allocation among the robots. Experimental results from different simulated and real environments that illustrate our approach for different communication conditions are presented.

Multi-robot repeated area coverage

“…Although there is a wide body of literature for single coverage scenarios (Choset 2001;Agmon et al 2008a;Rekleitis et al 2008;Gabriely and Rimon 2011;Zheng et al 2005;Kurabayashi et al 1996;Batalin and Sukhatme 2002), repeated coverage has not received the same attention. Two classes of the repeated coverage problem in the literature are: 1) Area Patrolling, and 2) Boundary or Perimeter Patrolling (Open or Closed Polylines), and each is divided into:…”

“…• Single Coverage: The aim is to cover the target area until all the accessible points of interest in the environment have been visited at least once, while minimizing the time, distance traversed by the robots, and the number of visits to the points (Fazli et al 2010a,b;Hazon and Kaminka 2008;Rekleitis et al 2008). …”

Exact Solution for Search-and-Rescue Path Planning

“…V c : maximum number of visits on cell c p cc : conditional probability of "correct" target detection on a visit in cell c given that the target is located in cell c. (14) Flow conservation…”

“…In counterpart, many contributions on search path planning may be found in the robotics literature in the area of robot motion planning [8] and, namely, terrain acquisition [9], [10] and coverage path planning [11]- [13]. Robot motion planning explored search path planning, primarily providing constrained shortest path type solutions for coverage problem instances [14], [15]. These studies typically examine uncertain search environment problems with limited prior domain knowledge, involving unknown sparsely distributed static targets and obstacles.…”

Multi-robot exploration and rendezvous on graphs

“…The multi-robot exploration task requires mobile robots to uniquely cover a given region of interest and also meet as quickly as possible to share their respective information. This process involves continuous communication and planning between the agents for coordination [2]. In contrast, we consider the case where the robots do not communicate during exploration and attempt to meet at regular intervals to find and exchange information with each other.…”