Commanding Heterogeneous Multi-Robot Teams

Bolstering up a military unit, e.g., an infantry platoon on a recce mission, by robots often is a double-edged sword. On the one hand, the robots are able to support the soldiers in multiple ways, they can transport bulky equipment, they can enter risky spots, and they may have sensor suits that help to detect dangers of all kinds. On the other hand, robots need to be equipped with energy sources that are cumbersome by themselves. In addition, the robots must be commanded. In order to optimize the support for their unit, it is often necessary to have a team of heterogeneous robots, e.g., UAVs as well as UGVs all with different sensors and specific abilities. Such a team, however, is even harder to command than a team of homogeneous robots. Our research aims at simplifying commanding teams of heterogeneous robots. In order to achieve this aim, we use the standards BML (Battle Management Language) and ROS (Robot Operating System) to communicate with the robot team. BML is used since our approach to commanding robots is from the language point of view very similar to commanding simulated units. Thus, we use language constructions modeled on those we developed as part of the NATO research groups on BML, NATO MSG-048 and NATO MSG-085. Currently, we are testing to use one single mobile GUI, also modeled on our NATO research groups’ results. That GUI is implemented on a tablet and enables the controller to command a team of two UAVs and four UGVs. All the robots can be equipped with different sensor suits. This article presents our solutions about the following topics essential for the described challenge. First, the robots have to introduce themselves by communicating their current abilities and their status to the commander. Second, the commander gives the commands to the robot team in a mission kind fashion. Thus, third, there has to be a kind of intelligence in the robot team that calculates sub tasks out of a given command and distributes these sub tasks to the team members taking the members’ specific abilities into account. Fourth, the robot team has to fuse sensor data in order to send a unified picture back to the commander in order to contribute to the operational picture.