Oblique projection based enumeration of mixed noncoherent and coherent narrowband signals

In many application scenarios, the signals impinging on the array of sensors comprise the noncoherent (uncorrelated and/or partially correlated) signals and the coherent signals with several groups due to multipath propagation. In this paper, we consider the problem of estimating the number of noncoherent narrowband signals and that of coherent narrowband signals with multiple groups impinging on a planar sensor array composed of two parallel uniform linear arrays (ULAs), and an oblique projection based enumerator for the mixed signals (OPEMS) is proposed by utilizing the QR decomposition based ratio criterion (QRRC) for rank determination of a matrix. In the OPEMS, the number of noncoherent signals and that of coherent signals in each group are estimated separately, where only the elevation angles of noncoherent signals are estimated to isolate the coherent signals from the noncoherent ones, and the computationally intensive and time-consuming eigendecomposition procedure is avoided. The consistency of the proposed OPEMS is analyzed, and its effectiveness is verified through numerical examples.