ANTENNA ARRAYS

The reduction of the errors in the observables caused by the multipath propagation and the interferences is addressed. Two di erent estimators are derived by applying the maximum likelihood (ML) principle to a signal model that assumes the reception of several re ections of the GNSS signal and that the spatial signatures of all the signals are arbitrary and unstructured. The rst estimator, which is a classical result since the noise is modeled as spatially white, is an extension of the well-known multipath estimating delay-lock-loop. The second estimator is derived by assuming that the spatial correlation matrix of the noise is unknown. This endows the estimator with interference cancelation capability, of which the rst estimator lacks. The second method constitutes a new result and its performance is always equal or better than that of the rst one. Moreover, we propose an approximation of the estimator for the correlated-noise case that provides the same performance as the original criterion. This approximation may allow the use of a computationally simple optimization algorithm that was only applicable in the white-noise case.