A Multi-robot Control System based on ROS for Exploring Disaster Environment

摘要

This control system is implemented to coordinate multiple robots in cooperative exploration tasks, in which a task assignment algorithm is used for calculating the expectation cost to achieve the required tasks and then allocate necessary tasks to specific robots. Robot control systems are used to gather crucial information about a disaster environment, such as a victim's position or changes in an indoor environment before a human is able to enter that environment. There are practical problems in using previous systems because these systems need professionals to operate these multi-robots. The proposed system automatically distributes tasks to each robot based on target positions that are selected by the person who does not possess any prior knowledge about robot operation. Therefore, the operator needs only to determine the target position of interest and not be concerned about controlling the robot. Moreover, the system is implemented using ROS in order to add functions and apply this system to various types of robots. This system has been verified and evaluated by Turtlebot 3 in the Gazebo simulator, as well as in a real setting by conducting the system usability scale (SUS) with 23 firefighters.