Sparse coarray manifold separation for efficient cellular localization using coprime array

Abstract

Advances in radio access network and antenna array processing have spurred the recent wave of research and trials into cost-effective schemes for cellular-based localization. To facilitate high-precision and low-latency position-based services, we propose a sparse coarray manifold separation (SCMS) method for fast joint direction-of-arrival and time-of-arrival estimation using a coprime array. By leveraging the Vandermonde structure in the manifold separation model, the two-dimensional (2D) spatial spectrum can be transformed into a discrete Fourier form and computed using the 2D robust random slice-based sparse Fourier transform. Through extensive numerical evaluations and link-level tests, we demonstrate that the SCMS method offers a precise approximation of true locations and significantly reduces computational complexity compared to baseline methods.