Robust Wideband 2-D DOA Estimation for Compact Uniform Circular Arrays in Shallow Water Doubly Spread Channels

Abstract

In shallow-water leader–follower AUV operations, the compact uniform circular array (UCA) carried by a micro autonomous underwater vehicle (AUV) suffers from a limited aperture. Together with strong coherent multipath and Doppler scaling, this limitation often causes high-resolution two-dimensional direction of arrival (DOA) estimation to break down. This paper proposes an STFT-RSS-DL-MVDR framework for wideband linear frequency modulation signals. The STFT restores the local validity of the narrowband array model by decomposing the nonstationary wideband signal into approximately stationary subband observations. RSS focusing alleviates the covariance rank under coherent multipath without sacrificing the physical aperture. The MVDR stage avoids source-number estimation, while diagonal loading regularizes the covariance inversion. Simulations using a compact five-element UCA with a radius of only 18.2 mm show that, under the tested strongly coherent multipath conditions, the proposed framework reduces the RMSE by approximately 40% relative to CSM-MUSIC at an SNR of −10 dB and exhibits substantially smaller degradation under path-dependent Doppler scaling. Robust estimation of the leader AUV direction is maintained when the number of paths approaches or exceeds the number of array elements and when the observation time is shortened to 5 ms. Accordingly, the framework provides a computationally efficient approach to 2-D target-DOA estimation for the tracked leader AUV within the modeled shallow-water doubly spread channel.