Beam Squint Assisted Joint Angle-Distance Localization for Near-Field Communications

The advent of extremely large-scale MIMO (XL-MIMO) provides crucial support for future sixth-generation (6G) systems to achieve high-precision, low-latency user localization. In near-field wideband scenarios, conventional beamforming based on frequency-independent phase shifters (PS) leads to a severe beam-squint phenomenon, where beams on different subcarriers point to inconsistent directions, thereby significantly degrading localization accuracy. However, by combining true-time-delay (TTD) elements and PS, the beam-squint effect can be exploited to establish a controllable mapping between subcarriers and spatial locations for localization purposes. Yet, most existing near-field localization methods adopt a two-step strategy that first estimates the angle and then the distance. This sequential process is susceptible to error propagation, impairing overall localization accuracy. To address these issues, in this paper we propose a beam-squint-assisted joint angle-distance (JAD) localization scheme with a coarse-to-fine two-stage estimator. The first stage rapidly obtains a coarse estimate of the user's position from the peak of the received power spectrum. The second stage applies a near-field improved multiple signal classification (MUSIC) algorithm to conduct a fine, high-resolution search around the coarse estimate, yielding a precise JAD estimate. Simulation results demonstrate that, compared with conventional two-step localization methods, the proposed method achieves significant improvements in localization accuracy and robustness in both single-user and multi-user near-field scenarios, highlighting its potential for future 6G sensing and localization systems.