An adaptive correction technique for Differential Global Positioning System

Accuracy and reliability are of the most important issues regarding the potential of the Global Positioning System (GPS) to become a sole means air navigation system. The purpose of this dissertation is to propose solutions to these problems.; The Differential GPS (DGPS) is the most effective technique to improve the accuracy of the GPS. Many experiment results prove that the DGPS can make significant improvements. However, the current DGPS accuracy still cannot meet the ICAO category II approach/landing precision requirements. Thus there is a need to explore the filtering technique to make further improvements.; Low pass filters, the Hatch/Eschenbach filter, and the Kalman filter are investigated as candidate filters to improve the quality of the correction signals; the Hatch/Eschenbach filter and the Kalman filter are implemented as the user filter to smooth navigation solutions. Better accuracy can be achieved by the optimization of the DGPS filter configuration. Through static and dynamic test it is found that the most accurate and robust choice for DGPS is the Hatch/Eschenbach filter. It was used as both reference station correction signal filter and user's navigation filter. Using this Hatch/Eschenbach filter configuration, the accuracy under the elevation mask of 7.5{dollar}spcirc{dollar} is: lateral--0.98 m (2{dollar}sigma{dollar}); vertical--1.81 m (2{dollar}sigma{dollar}); absolute positioning--2.50 m (2{dollar}sigma{dollar}).; One serious problem of DGPS is the latency effect which is the accuracy degradation caused by the delay of the correction signals. It greatly reduces the DGPS system reliability. To alleviate the latency effect, an adaptive correction technique (ACT) is developed. This technique adaptively identifies the autoregressive model parameters and order of the correction signals by using a Recursive Least Squares Lattice Filter (RLSLF). Using these parameters, the {dollar}Delta{dollar}-step predictor of the RLSLF can generate better correction signals than the existing RTCM SC-104 correction algorithm. The static test results proves that ACT is better than RTCM SC-104 when the latency is longer than 180 seconds. In addition, it is very effective to suppress the growth of positioning error for very long latency.; Based on results of the current research and the GPS 21 primary satellite constellation, a DGPS-aided landing precision database is constructed. This database provides the landing precision information of various combination of latitudes and mask angles. It is a projection of the landing precision of the DPGS implementation in this research under full operational GPS.