UNILocPro: Unified Localization Integrating Model-Based Geometry and Channel Charting

TL;DR

The paper tackles high-precision localization in GNSS-degraded environments by integrating model-based geometry with channel charting in a unified, unsupervised framework (UNILocPro). It introduces novel dissimilarity metrics based on gospa and Wasserstein distances, plus an OT-based map-alignment loss to preserve global geometry, and a low-complexity variant UNILoc that reduces training cost by pre- or one-shot OT steps. The approach yields significant performance gains over purely model-based or CC-based methods, with UNILocPro plus timestamps approaching fully supervised fingerprinting despite no training labels. This framework provides robust, scalable localization for mixed LoS/NLoS scenarios and enables practical deployment without extensive labeled data or precise loss/nlos identification.

Abstract

In this paper, we propose a unified localization framework (called UNILocPro) that integrates model-based localization and channel charting (CC) for mixed line-of-sight (LoS)/non-line-of-sight (NLoS) scenarios. Specifically, based on LoS/NLoS identification, an adaptive activation between the model-based and CC-based methods is conducted. Aiming for unsupervised learning, information obtained from the model-based method is utilized to train the CC model, where a pairwise distance loss (involving a new dissimilarity metric design), a triplet loss (if timestamps are available), a LoS-based loss, and an optimal transport (OT)-based loss are jointly employed such that the global geometry can be well preserved. To reduce the training complexity of UNILocPro, we propose a low-complexity implementation (called UNILoc), where the CC model is trained with self-generated labels produced by a single pre-training OT transformation, which avoids iterative Sinkhorn updates involved in the OT-based loss computation. Extensive numerical experiments demonstrate that the proposed unified frameworks achieve significantly improved positioning accuracy compared to both model-based and CC-based methods. Notably, UNILocPro with timestamps attains performance on par with fully-supervised fingerprinting despite operating without labelled training data. It is also shown that the low-complexity UNILoc can substantially reduce training complexity with only marginal performance degradation.

Figures (11)

• Figure 1: A street canyon scenario sionna, where channel paths are shown for user 1 for example. Note that user 1 is a los user, while user 2 is a nlos user as its los path is blocked by a building.
• Figure 2: The framework of UNILocPro (the nn training process is shown in Fig. \ref{['fig:UNILocPro_training']}) and UNILoc (the nn training process is shown in Fig. \ref{['fig:UNILoc_training']}).
• Figure 3: The training process of the nn model in UNILocPro (the red dashed arrow line indicates the path of gradient backpropagation).
• Figure 4: The computed dissimilarity v.s. the physical distance between user positions, where the red line represents the real distance (using the same system setting as described in Sec. \ref{['sec:system_setting']}).
• Figure 5: The training process of the nn model in UNILoc (the red dashed arrow line indicates the path of gradient backpropagation).

Theorems & Definitions (2)

• Remark
• Remark