| Popis: |
Fuzzy Cognitive Maps (FCMs) provide a powerful framework for developing supervisory, mission controllers which we term Fuzzy Cognitive Controllers (FCCs). In an FCC, control concepts such as actions, events, and action motivators are represented as nodes connected in a causal web. Autonomy comes from the interplay between these concepts and is encoded in the overall architecture of the web along with the weighted influences of each node on all the other nodes. This paper presents the basic concepts of Fuzzy Cognitive Controllers, and then illustrates them using the development of a Fuzzy Cognitive Controller for a notional UAV with autonomous behavior capabilities. The FCC governs all aspects of the vehicle’s behavior including its trajectory, sensor payload manipulation, and its interface with an operator. The UAV behaves in an intelligent, autonomous manner by monitoring its payload sensors and altering its flight behaviors in response to sensor inputs as well as operator commands. The operator interacts with the UAV at a very high level through linguistic-based commands to the individual nodes representing system’s various high level modes of behavior. Results are presented showing the FCC’s control of a small UAV during performance of a notional mission. Preliminary results indicate that FCCs offer a practical approach to developing UAV mission controllers that can drastically reduce an operator’s workload. The technique also lends itself naturally to the control of multiple vehicles, which we shall present in a future paper. |
