Stochastic strategies for patrolling a terrain with a synchronized multi-robot system

TL;DR

This paper introduces stochastic patrolling strategies for synchronized multi-robot systems, demonstrating through theoretical analysis and simulations that these methods outperform deterministic approaches in coverage and communication metrics.

Contribution

It provides a novel theoretical framework for analyzing stochastic patrolling strategies and empirically shows their advantages over existing deterministic methods.

Findings

Stochastic strategies reduce idle and isolation times compared to deterministic methods.

Theoretical results align well with simulation outcomes.

Random strategies improve broadcast time in multi-robot patrolling.

Abstract

A group of cooperative aerial robots can be deployed to efficiently patrol a terrain, in which each robot flies around an assigned area and shares information with the neighbors periodically in order to protect or supervise it. To ensure robustness, previous works on these synchronized systems propose sending a robot to the neighboring area in case it detects a failure. In order to deal with unpredictability and to improve on the efficiency in the deterministic patrolling scheme, this paper proposes random strategies to cover the areas distributed among the agents. First, a theoretical study of the stochastic process is addressed in this paper for two metrics: the \emph{idle time}, the expected time between two consecutive observations of any point of the terrain and the \emph{isolation time}, the expected time that a robot is without communication with any other robot. After that, the random strategies are experimentally compared with the deterministic strategy adding another metric: the \emph{broadcast time}, the expected time elapsed from the moment a robot emits a message until it is received by all the other robots of the team. The simulations show that theoretical results are in good agreement with the simulations and the random strategies outperform the behavior obtained with the deterministic protocol proposed in the literature.