Coalition of Operators and Automation for Collaborative Tasking (COACT)

Increasing the autonomy of unmanned vehicles potentially reduces operator workload and cognitive load. Realizing the promise of autonomy however, is difficult when deploying in environments with mobility and visibility challenges, enemy threats, risks of endangering civilians, and limited communication. Achieving full autonomy in the face of these challenges is both a hard technical problem and does not take advantage of valuable human capability. Responsibility to achieve mission goals should be shared between humans and autonomous vehicles to improve mission success and timeliness. TRACLabs will develop techniques and software to improve interaction and collaboration among teams of humans and autonomous vehicles. We will use a common semantic model (ontology) to ground interaction about mission goals and resources, as well as vehicle capabilities and constraints. This shared understanding should improve the transparency of autonomous behaviors. We will improve remote presence by using the ontology to develop a shared perception of the mission. We will use multi-agent procedures to coordinate the activities of distributed operators and autonomous vehicles, and to shift task responsibility among these agents during the mission. The resulting transparency of autonomous vehicle behavior and ability to flexibly share tasks among humans and autonomous vehicles should improve trust in vehicle autonomy.