Near-Field LoS MIMO Channel Representation Based on Consecutively Pairwise-Division for Phase Linearization

초록

In this paper, we propose a new channel representation method that facilitates a user localization problem in near-field line-of-sight (LoS) channel environments. Our approach is to harness a phase linearization method. Unlike a far-field channel model, the phase response of the near-field channel is characterized as a quadratic function across antennas, which arises from the second-order Taylor approximation of the spherical wavefront. Thus, in a continuous domain, the differentiation of this function yields a linear function, which forms the basis of our method. To implement this in the spatial sampling domain, pairwise division relationships are established between the elements of the observed channel. We refer to this near-field LoS MIMO channel representation based on consecutively pairwise division (CPD). More specifically, there are two representative types of patterns in terms of the index pairing pattern. One pattern maintains a constant difference between indices, whereas the other maintains a constant sum. In order to validate our proposed CPD method, we demonstrate through simulations that our approach enables user localization with significantly improved accuracy compared to conventional methods. © 2026 Institute of Electrical and Electronics Engineers Inc.. All rights reserved.

키워드