Using Continuous-Planning Techniques to Achieve Autonomy and Coordination Among Multiple Unmanned Aerial Vehicles

摘要

Existing Unmanned Aerial Vehicle (UAV) systems exhibit shortcomings in providing continuous, responsive, timely, and detailed information and targeting support to Army tactical commander's combat operations in an Army XXI battlespace. To synchronize Tactical UAV (TUAV) missions with supported operations, time is the critical element. Anything that can reduce TUAV planning time, while maintaining plan effectiveness, will expedite execution of a TUAV1s mission. Autonomous flight with some ability to avoid and evade certain threats would increase survivability further. This paper begins by presenting some of the recent successes achieved by artificial intelligence (Al) planners and schedulers on complex real-world problems. It then attempts to show how NASA's demonstrated utility of a dynamic Al planning system prototype for conducting autonomous distributed planning and execution for a team of rovers engaged in missions to achieve science goals during planetary operations can be generalized and applied to a team of TUAVs. Last, it discusses some data collection opportunities that should appear due to the ability to place increasingly more processing and data storage capabilities onboard TUAVs and some of the key challenges to use those capabilities to produce more timely and immediately usable interpretations.