A multi-agent simulator environment based on the robot operating system for human-robot interaction applications

Abstract

This paper describes a simulator environment for humans to direct a team of independent drones by allowing humans to issue high-level goals to the teams or drones. Given a goal, the environment will generate plans for the drones and monitor their execution while attending to humans requests (e.g., aborting a goal, introducing a new goal). For this reason, the environment includes two specific modules, a planning module and an execution and monitoring module, besides the modules for simulation and control of drones. The environment is implemented on the Robot Operating System (ROS), a well-known framework for the development of robotic applications, that facilitates the communication between its components. Experiments are included to highlight the applicability of the environment.