Map-Aided Particle Filter for Improved Multi-hypothesis Ambiguity Resolution

This paper presents the implementation of a map-aiding algorithm for estimating the full geometry-based float solution in the differential carrier phase positioning of a land vehicle. The focus is on evaluating the effect of map-matching on the convergence of ambiguity states. The estimation is performed within a Bayesian framework using a Sequential Importance Resampling (SIR) Particle Filter (PF). The SIR PF estimates the position, velocity, acceleration states, and float ambiguities using L1 GPS carrier phase and pseudorange observations. Three ambiguity resolution methods, namely Rounding, Bootstrapping, and Integer LeastSquares (ILS), are applied to determine the most likely integer values based on the estimated float ambiguities. The empirical covariance matrices for ambiguity resolution are obtained by analyzing the particle distribution with and without the map matching and transformed using the decorrelation Z transformation of the Least Squares Ambiguity Decorrelation Adjustment (LAMBDA) method. The results demonstrate the benefits of incorporating map information in the ambiguity resolution process, particularly in improving ambiguity convergence. The success rates of the three search methods are computed and compared to show the positive effect of the map-aiding algorithm on the ambiguity domain. The proposed algorithm is tested on 4 segments of data from a land vehicle test chosen to demonstrate ambiguity re-initialization and convergence during different vehicle maneuvers with and without map-aiding.