Development of a Low-Latency, High Data Rate, Differential GPS Relative Positioning System for UAV Formation Flight Control

摘要

In order for Unmanned Aerial Vehicles (UAVs) to be able to fly missions currently performed by manned aircraft, they must be able to conduct in-flight refueling. Additionally, significant fuel savings can be realized if multiple UAV's are able to fly in precise formation and align wingtip vortices. In either case. the precise relative position between the aircraft must be known to an accuracy of only a few centimeters. Previous research at the Air Force Institute of Technology culminated in the development of a relative positioning system for manned aircraft. This thesis presents the development of the next-generation system designed for small UAV's. Because of the stringent size, weight, and power consumption requirements inherent in small UAV's, several approaches were taken to maximize efficiency and performance while simultaneously keeping the system small and lightweight. At the core of the Differential GPS (DGPS) application presented in this thesis are three separate asks which operate asynchronously yet share information when required. A Kalman filter task operates continuously at a 1 Hertz rate. An ambiguity resolution task, utilizing the Least squares Ambiguity Decorrelation Adjustment (LAMBDA) method, is run whenever the floating point ambiguities must be resolved to their integer values. A high-rate output task, operating at a 20 Hertz rate formulates a high-rate, centimeter-level, relative position solution with less than 10 milliseconds of latency. The use of wide lane measurements generally resulted in a 2 second convergence time for ambiguity resolution and a 99.9 percent success rate of selecting the proper ambiguity set. However, in order to minimize the increased errors associated with multipath, the system quickly transitions from wide lane mode to narrow lane mode. The system was tested on the ground in both a static and dynamic environment.