Stochastic-deterministic multipath model for time-delay estimation

Line-of-sight (LOS) delay estimation in multipath scenarios is a central problem in global navigation satellite systems (GNSS). Deterministic channel models can be used to describe the multipath environment, but this usually requires the estimation of several nuisance parameters. In order to avoid this effort, stochastic channel models can be used. In this case the multipath statistics have to be estimated. The correlated path (CP) model combines the two approaches, dividing the multipath signal into a part correlated with the LOS signal, and the rest as uncorrelated multipath interference. In an earlier version of the CP model the multipath interference was modeled as temporally white noise. This is not an accurate assumption however, especially in the case when a bank of correlators is used for signal compression. In this paper we derive the temporal covariance matrix of the multipath interference and show how the temporal multipath correlation can incorporated into the maximum-likelihood (ML) estimator of the CP model. Simulation results show that this new approach leads to better LOS time-delay estimation performance.