Introduction
Today, computers with significant computational capabilities continually become less expensive. The economics of consumer electronics has also made inexpensive GPS units available. A remarkable notion is to develop a guidance and flight control architecture that utilizes a single antenna GPS as a sole sensor 1, resulting in a very low cost system, still capable of rudimentary waypoint navigation, altitude control, and approach courses. However, when developing a flight control system for an Uninhabited Air Vehicle (UAV) using GPS alone, several issues arise. These inexpensive commercial units typically have update rates of 1 Hz and significant latency. Also, position and velocity is less information than that conventionally used for flight control, where air data and gyroscopic instruments are commonplace. The challenge is develop algorithms for navigation and flight control which maximize the robustness, tracking bandwidth, and accuracy given this limited sensor capability. An architecture is proposed, which utilizes observers for bank angle and wind components to reconstruct this “missing” sensor information. The airplane is assumed to be statically stable, and with a short-period mode that is fast compared to the response of the autopilot. By these assumptions, angle of attack is effectively set directly by elevator position at a given airspeed. This architecture has been implemented on a small fixed-wing UAV shown figure 1. Low Cost Aerial Vehicle, with Avionics Mounted Above Wing (GPS Receiver), and in Boxes on the Sides of the Fuselage (Computer and Battery)