Fusion of Optimal Control and Agent Based Reasoning for Unmanned Aerial Vehicles
Galen Brambley, Michael Matkovic · Infotech@Aerospace · 2005
A two-tiered igh t control system consisting of an agent with an H1 optimal controller is presented, with results obtained from a xed time step igh t simulator. System performance is monitored using the k k1 norm and when deemed unsatisfactory, alerts the agent. The agent categorizes the drop in performance into one of three possible causes, environmental, platform related and mismatch of aircraft to controller. Remedial actions available to the agent include mission adjustment, a reduction in performance, recalculation of the controller using coecien ts extracted from the simulator during run time. Methods of deliberation are presented along with the methodology behind the cooperative control. Preliminary results show an acceptable level of fault detection in the agent system.