Drone Fisher

Autonomous agents are becoming more prevalent in a variety of domains as synthetic teammates for training one or two individuals in team tasks. Where instructors used to play the complicated roles of the teammates as well as “game master” the scenario, the inclusion of synthetic teammate role players in training environments is alleviating the pressure on the instructors by allowing them to focus on the instruction and feedback rather than also playing the game realistically for the students. This requires the instructors to become familiar with how to maintain situational awareness for effective instruction and feedback when the autonomous agents don’t always communicate the intricacies of their decisions, however the human student must react accordingly. DroneFisher allows users who may be working with autonomous agents for task performance to practice with the technology and learn how to use speech for commands, understand how to query and track autonomous updates, and the effective use of autonomous agents. The domain is agnostic for reusability across platforms and domains, with a simple game that is less about the content being trained and specific agent behavior and is instead focused on learning to collaborate with autonomous agents to complete a task. This work is funded by the Naval Air Systems (NAVAIR) Naval Air Warfare Center Training Systems Division (NAWCTSD) via a Phase II Small Business Innovative Research (SBIR) award, and is allowing the Navy to conduct automation transparency research in order to improve the design for autonomous agent feedback and interfaces in team training paradigms.