Human-Interaction Challenges in UAV-Based Autonomous Surveillance

Abstract Autonomous UAVs provide a platform for intelligent surveillance in application domains ranging from security and military operations to scientific information gathering and land management. Surveillance tasks are often long duration, requiring that any approach be adaptive to changes in the environment or user needs. We descriie a decision- theoretic model of surveillance, appropriate for use on our autonomous helicopter, that provides a basis for optimizing the value of information returned by the UAV. From this approach arise a range of challenges in making this iiamework practical for use by human operators lacking specialized knowledge of autonomy and mathematics. This paper describes our platform and approach, then descriies human-interaction khallenges arising hn this approach that we have .I identified and begun to address. UAV-based Surveillance One of the earliest applications of powered air vehicles was to gather information about conditions on the ground, exploiting the relatively high speed and broad view offered by these machines to, e.g., provide guidance to World War I artillery units and track enemy movements. Similar applications quickly emerged in other areas such as security, land management and scientific research. Unmanned aerial vehicles (UAVs) have the potential to dramatically increase the availability and usefulness of aircraft