HRILaboratory's video: Task-Based Multi-Robot Dialogs UGV

In this multi-robot scenario, both robots are controlled by the DIARC robot architecture. We are able to task the robots using spoken natural language commands, and the robots maintain belief models of each of the other agents in the scenario. The mule's models allow it to infer that its destination is the same as the transport's, despite having been told only to follow the transport. Similarly, the transport detects that the commander believed incorrectly that it had a goal to be at alpha and takes steps to correct the problem. DIARC is implemented in the ADE robot development framework, which allows control architectures to be specified without detailed knowledge of the underlying hardware. This allows us to substitute an airborne drone (http://www.youtube.com/watch?v=40_Ee2g5ztg) for the ground vehicle without making any changes besides switching the component that sends motor commands to the robot. For additional information, please visit http://hrilab.cs.tufts.edu/.