The complete integration of MissionLab and CARMEN

摘要

Nowadays, a major challenge in the development of advanced robotic systems is the creation of complex missions for groups of robots, with two main restrictions: complex programming activities not needed and the mission configuration time should be short (e.g. Urban Search And Rescue). With these ideas in mind, we analysed several robotic development environments, such as Robot Operating System (ROS), Open Robot Control Software (OROCOS), MissionLab, Carnegie Mellon Robot Navigation Toolkit (CARMEN) and Player/Stage, which are helpful when creating autonomous robots. MissionLab provides high-level features (automatic mission creation, code generation) and a graphical mission editor that are unavailable in other significant robotic development environments. It has however some weaknesses regarding its map-based capabilities. Creating, managing and taking advantage of maps for localization and navigation tasks are among CARMEN's most significant features. This fact makes the integration of MissionLab with CARMEN both possible and interesting. This article describes the resulting robotic development environment, which makes it possible to work with several robots, and makes use of their map-based navigation capabilities. It will be shown that the proposed platform solves the proposed goal, that is, it simplifies the programmer's job when developing control software for robot teams, and it further facilitates multi-robot deployment task in mission-critical situations.