18th ICCRTS: C2 in an Underdeveloped and/or Denied Environment

The DoD’s envisioned role shift for humans from operators to future autonomous systems supervisors presents significant challenges for developing effective decision support. What decision support will supervisors need to effectively oversee autonomous systems? What will their task needs be and how can we support them with usable and useful supervisory human-machine interfaces (HMIs) and tools? Here, we inform the requirements and design of future decision support through a systematic cognitive engineering and analysis process. Structured interviews with 27 unmanned systems experts were carefully sequenced across four groups, with results and artifacts from one group informing the next interviews. Interviews focused on supervisory monitoring and intervention tasks and were designed to feed a userand task-centered, scientifically-principled HMI design process to develop the decision support. The interviews informed this design process by populating three key design artifacts: (1) a model of current and future tasks, (2) their allocation across humans and automation, and (3) the necessary supporting human-automation exchanges. A framework for system designers to allocate tasks across humans and automation was extended to provide an objective basis for subject matter experts to contribute their expectations for future automation. The interview results show how today’s task needs are not met by current HMIs and tools, and how persisting with them is unlikely to meet the future needs of more nuanced supervisory decision making. Our results inform the design of future supervisory HMIs that target and mitigate today’s capability gaps and shed light on how to begin to achieve the DoD vision.