DOA estimation and association of coherent multipaths by using reference signals

This paper deals with multiuser localization by array processing in the case of multipath propagation. A high-resolution (HR) method based on the source subspace determination is used to estimate the directions of arrival (DOA) on a multiple antenna array. A parallel and separate processing of the different users is proposed by using reference signals and thus a natural association between the estimated parameters and the transmitters is achieved and the constraint on the antenna size is relaxed. Indeed, a reference signal or equivalently the known waveform of each transmitting source, supposed orthogonal to the others, is assumed to be available at the receiver during the processing period. The reference signal is used to reject the noncoherent interference and to estimate the spatial signature of the signal or equivalently the corresponding channel response. Then an appropriate smoothing is applied and DOA of each user paths are separately recovered. The so-called PADEC parallel decomposition algorithm exploits the shift invariance properties of linear and uniform antenna arrays. To assess the derived algorithm capability to handle coherent multipaths and to examine its estimation accuracy, some comparisons with a variety of maximum likelihood (NIL) and subspace algorithms, that either exploit the known reference signals or not, are achieved. The asymptotic analysis of PADEC is also presented. (C) 2004 Elsevier B.V. All rights reserved.