Fuzzy Logic Approach in Real-time UAV Control

Abstract

The current article presents a fuzzy logic approach for creating an artificial intelligent and autonomous pilot for unmanned aerial vehicle. The article presents a new way of identifying the position of the UAV relative to a destination point, based on two angles: deflection angle and relative angle. The major advantage of the presented method (over other artificial intelligent methods like genetic programming or neural networks) is the way of taking decisions in real-time, without the need of presenting any domain specific data to the system before running the simulation. In this article, a professional flight simulator - Microsoft Flight Simulator X (FSX) - is used and its software development kit (SDK), and therefore, all computations are specific to this flight simulator. The aircraft that is used is a real UAV - General Atomics RQ/MQ-1 Predator - an FSX add-on.