Reinforcement Learning and Genetic Learning Classifier Systems for Sensor Management and Adaptive Flight Control System

The work proposed here will build on current research results to develop and implement algorithms for planning trajectories of multiple Unmanned Aerial Vehicles (UAVs) working cooperatively. Their common mission is to reach a target by flying in an unknownenvironment, which they learn about through onboard sensors. The resulting sensor resource control problem is solved via dynamic programming. The algorithms to be developed and flight tested are envisioned as part of a first-of-a-kind autonomous VTOL UAVsystem being developed by NTC. This system will bring military operations closer to the goal of safe, effective operation in dangerous urban warfare environments. In addition, NTC is pursuing commercial uses of UAVs in education, newscasting, andentertainment.