Adaptive Levels of Autonomy (ALOA) for UAV Supervisory Control

Abstract : An architecture tor testing and evaluating different methods for adaptive levels of autonomy was devised. We defined multiple Levels of Autonomy (LOA) for each of four operator tasks: allocation, route planning, imagery analysis, and weapon control. To demonstrate the architecture and LOA implementation, we designed a prototype Multi-UAV Control Station Emulator research test bed, by building on existing ORCA-developed software components. ORCA's extensive internal IR&D over several years has produced state-of-the-art automated mission planning tools that allow fully autonomous execution of operator tasks. Experience with operators through the J-UCAS effort and other programs gives us first-hand knowledge of the tools and decision aids operators need when building and assessing mission plans, which support manual mission planning. With this experience, we implemented the two autonomy extremes: manual and fully autonomous and we defined and implemented intermediate levels of autonomy (requires using characteristics of both manual and autonomous task execution for the four operator tasks noted above.