Computationally efficient method for estimation of the number of signals in multipath environment

This paper proposes a nonparametric detection method for the coherent narrowband signals impinging on a uniform linear array (ULA). By exploiting the array geometry and its shift invariance property, the coherency of incident signal is decorrelated through subarray averaging, and the number of incident signals is equal to the rank of a matrix formed from the cross-correlations between some sensor data, where the effect of additive noise is also eliminated. Then a new criterion is formulated in the terms of the row elements of the QR upper-triangular factor of the cross product of formed correlation matrix, and the number of signals is determined as the value for which this QR-based criterion is maximized. The proposed method has remarkable insensitivity to the correlation of incident signals and flexibility to the spatially correlated noise. Simulation results show that the proposed method is superior in detecting closely-spaced signals with small number of snapshots and at low signal-to-noise ratio (SNR).