A Co-Array Approach to Multi-Source 2D-DOA Estimation with Time-Multiplexed Array Elements

This paper addresses the problem of multi-source direction finding with a system that employs fewer receivers than sensors. We propose a new system architecture, specifying a time-multiplex scheme together with a refined array geometry. The proposed array geometry consists of a centrosymmetric prototype array that is connected to the receiver through a switch and multiple reference elements that are directly connected to the receiver. Under the assumption of wide-sense stationary source signals, it is possible to reconstruct the array signals of the prototype array, as if it was sampled simultaneously. The resulting signals are amenable to conventional array processing techniques and direction of arrival (DOA) estimators such as the MUltiple SIgnal Classification (MUSIC) method. The processing steps exploit the geometry of co-arrays constructed at each switch position relating to the time-multiplex scheme. The problem of fully coherent equivalent source signals, inherent to the co-array approach, is mitigated by spatial smoothing. Finally, the proposed technique is verified using numerical experiments and compared to the Cramer-Rao lower bound (CRLB).