An Unmanned Aerial Vehicle Controller based on a Learning Classifier System

Autonomous Unmanned Aerial Vehicle (UAVs) have been increasingly employed by researchers, commercial organizations and the military to perform a variety of missions. This thesis discusses the design of an autonomous controller using a Learning Fuzzy Classifier System (LFCS) to store and evolve fuzzy rules and fuzzy membership functions. The controller executes the fuzzy inference process and assigns credit to the population during a flight simulation. This framework is useful in evolving a sophisticated set of rules for the controller of a UAV, which deals with uncertainty in both its internal state and external environment. A flight simulation is implemented in Matlab/Simulink providing the opportunity to assess the accuracy of the control rules. The simulation results show that this approach is able to develop a controller that achieves high effectiveness in both lateral and longitudinal control.