Multi-Robot Frontier Based Map Coverage Using the ROS Environment

Abstract

Cooperative robotics deals with multiple robot platforms working to accomplish a common goal and has a multitude of applications including security, surveying, search and rescue, and many more. The use of multiple robots allows a task to be completed more efficiently, and is less prone to failure in the event that one of the robots becomes immobile. The Robotic Operating System (ROS) is a mainstream software framework being used for robotic research around the world. Despite its popularity and the very strong robotic community behind its success, there has not been much work with ROS involving multirobot teams. This thesis presents a complete multi-robot system implemented within the ROS software framework. Specifically, this work implements a collaborative robotic system that performs map coverage of a known environment. A team of robots is designed and programmed to cover a map with their range sensors. A list of frontiers that border searched and unsearched space is maintained. Each robot is assigned to travel towards unsearched space until the entire map has been covered. The frontier-based coverage method is evaluated through a series of simulation experiments in which the coverage planner is tested in different map environments while varying the number of robots in the system. The ROS-based implementation of multi-robot frontier coverage is shown to successfully be able to cover an entire area with a team of autonomous robots.