Research issues in autonomous control of tactical UAVs

This paper summarizes the enabling technologies for an autonomous tactical unmanned aerial vehicle (UAV). Current technologies are adequate for semi-autonomous UAVs that operate in a relatively structured environment. For tactical UAVs in a rapidly changing uncertain environment the present techniques are inadequate. The essence of autonomous control is rapid in-flight replanning under uncertainty. This is cast as a large optimization or decision problem. Monolithic and decomposition techniques are discussed for the solution of large decision problems. Hierarchical decomposition is the most promising approach, but suffers from an inadequate theoretical basis. Finally, research areas are proposed to address the decomposition problem.