Application of Least Square Method with Variable Parameters for GPS Accuracy Improvement

Abstract

In this work research results of Least Square Method for information filtering with sliding window are presented. Length of window changes in filtering process and depends on the evaluation results between measuring and expected coordinate. Modeling results of filtering with variable parameter window are comparing with constant length sliding window filtering. Algorithms for sliding window change are developed and research. Modeling results using variable sliding window coordinate estimation show that estimation error decrease about 10 time if object movement parameters are constant and decrease about 3 times relative measuring error if parameters change very rapidly. Results of sliding filter modeling and filter parameter optimization and use of optimal sliding window filter for GPS receiver information processing are described.