A Bayesian method for GPS signals delay estimation

In GPS applications, positioning techniques are based on the characteristics of the pseudo-random code autocorrelation function. They don't take into account the eventuality of multipath propagation. One reflected signal may greatly disturb the measures of the usual Early-Late method. A maximum a posteriori (MAP) estimator is proposed in this study, introducing a multi-path model and prior laws on both amplitude and delay parameters. This method deals with deconvolution of known functions with simple forms. Performance depends on data observation length but they greatly improve results.