Direction-of-arrival estimation in closely distributed array exploiting mixed-precision covariance matrices

Yimin D. Zhang,Waqeeb T.S. Chowdhury,Md Waqeeb T.S. Chowdhury

DOI: https://doi.org/10.1016/j.sigpro.2024.109463

IF: 4.729

2024-03-14

Signal Processing

Abstract:In this paper, we explore a collaborative direction-of-arrival (DOA) estimation technique that utilizes multiple closely spaced subarrays to maximize the potential of distributed arrays while minimizing communication overhead between the subarrays and the processing center. Each subarray computes its self-covariance matrix using the full-precision data and transmits it, along with a one-bit version of the measured data, to the processing center. The processing center generates one-bit cross-covariance matrices between subarrays, which are combined with full-precision subarray self-covariance matrices to create the mixed-precision covariance matrix of the entire array for source DOA estimation. This approach utilizes the full array aperture and all available degrees of freedom of the entire distributed array. To address missing entries in the full covariance matrix, we employ matrix completion, taking into account its Toeplitz and Hermitian structure. For subarrays that are not positioned on the half-wavelength grid, we propose an iterative DOA estimation method to ensure robust DOA estimation performance. Our proposed approach outperforms scenarios where cross-covariance matrices are unavailable or the entire covariance matrix is not interpolated. With the same communication traffic limitation, it demonstrates superiority over schemes that utilize only full-precision data or only one-bit data.

engineering, electrical & electronic