Adaptive Downlink Localization and User Tracking in Near-Field and Far-Field: A Trade-Off Analysis

TL;DR

The paper studies adaptive downlink localization and UE tracking across near-field and far-field regimes, introducing two signaling schemes (FF-based and NF-based) and two corresponding localization algorithms that leverage distinct beam design principles. It shows that a memory-enabled tracking framework can balance accuracy and computational/ signaling costs by transitioning between FF and NF modes based on UE distance and trajectory information, revealing that the Fraunhofer distance alone is an inadequate boundary in mixed NF/FF environments. The FF approach offers low-complexity localization at larger ranges, while the NF approach provides higher accuracy at the expense of greater signaling and computation; an adaptive protocol orchestrates switching to preserve overall performance. Simulation results confirm that the proposed adaptive scheme can approach NF-level accuracy with reduced overhead and latency, while the tracking memory enhances reliability and reduces unnecessary switching. Overall, the work demonstrates a practical, adaptive framework for DL localization and tracking in mixed NF/FF scenarios with explicit trade-offs between accuracy, signaling, and complexity.

Abstract

This paper considers the problem of downlink localization and user equipments (UEs) tracking with an adaptive procedure for a range of distances. We provide the base station (BS) with two signaling schemes and the UEs with two localization algorithms, assuming far-field (FF) and near-field (NF) conditions, respectively. The proposed schemes employ different beam-sweep patterns, where their compatibility depends on the UE range. Consequently, the FF-NF distinction transcends the traditional definition. Our proposed NF scheme requires beam-focusing on specific spots and more transmissions are required to sweep the area. Instead, the FF scheme assumes distant UEs, and fewer beams are sufficient. We derive a low-complexity algorithm that exploits the FF channel model and highlight its practical benefits and the limitations. Also, we propose an iterative adaptive procedure, where the signaling scheme is depends on the expected accuracy-complexity trade-off. Multiple iterations introduce a tracking application, where the formed trajectory dictates the validity of our assumptions. Moreover, the range from the BS, where the FF signaling scheme can be successfully employed, is investigated. We show that the conventional Fraunhofer distance is not sufficient for adaptive localization and tracking algorithms in the mixed NF and FF environment.

Figures (8)

• Figure 1: Considered scenario where BS provides 2 signaling options, designed for DL UE localization in the FF and the NF, respectively. The UE may employ a localization algorithm fit for the corresponding localization scheme.
• Figure 2: Proposed DL localization signaling procedure.
• Figure 3: Normalised relative complexity for different signaling overhead requirements for the two proposed Algorithms.
• Figure 4: Normalised compatibility metric for different UE distributions and different array sizes. Solid lines refer to $N_{\rm{BS}}=24\times 24$ and dashed lines to $N_{\rm{BS}}=16\times 16$.
• Figure 5: Position RMSE for the two proposed signaling schemes. The shaded region marks the possible $d_{\rm{NF}}$. The dashed lines mark the corresponding PEBs. Scenario A is considered.