A ROS-Based Framework for Simulation and Benchmarking of Multi-robot Patrolling Algorithms

Abstract

Experiments with teams of mobile robots in the physical world often represent a challenging task due to the complexity involved. One has to make sure that the robot hardware configuration, the software integration and the interaction with the environment is thoroughly tested so that the deployment of robot teams runs smoothly. This usually requires long preparation time for experiments and takes the focus away from what is essential, i.e. the cooperative task performed by the robots. In this work, we present patrolling_sim, a ROS-based framework for simulation and benchmarking of multi-robot patrolling algorithms. Making use of Stage, a multi-robot simulator, we provide tools for running, comparing, analyzing and integrating new algorithms for multi-robot patrolling. With this framework, roboticists can primarily focus on the specific challenges within robotic collaborative missions, run exhaustive tests in different scenarios and with different team sizes in a fairly realistic environment, and ultimately execute quicker experiments in the real world by mimicking the setting up of simulated experiments.