Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Spatially-Adaptive Conformal Graph Transformer for Indoor Localization in Wi-Fi Driven Networks

Ayesh Abu Lehyeh, Anastassia Gharib, Safwan Wshah

TL;DR

The paper tackles indoor localization with a focus on reliability by introducing SAC-GT, which fuses a Graph Transformer for precise 2D position estimates with Spatially-Adaptive Conformal Prediction to produce region-specific, statistically valid confidence regions. The GT captures spatial topology and RSSI dynamics, while SACP calibrates adaptive radii per region using a held-out calibration set, enabling tailored uncertainty that mirrors environmental heterogeneity. On the SODIndoorLoc benchmark, SAC-GT delivers a median localization error of $1.37$ m and strong 95th-percentile performance, plus an overall predictive coverage of $84.8\%$ against a $90\%$ target, illustrating robust real-world applicability. The work advances practical indoor localization by providing both high accuracy and reliable uncertainty guarantees, with potential impact on navigation, asset tracking, and safety-critical applications in Wi-Fi driven networks.

Abstract

Indoor localization is a critical enabler for a wide range of location-based services in smart environments, including navigation, asset tracking, and safety-critical applications. Recent graph-based models leverage spatial relationships between Wire-less Fidelity (Wi-Fi) Access Points (APs) and devices, offering finer localization granularity, but fall short in quantifying prediction uncertainty, a key requirement for real-world deployment. In this paper, we propose Spatially-Adaptive Conformal Graph Transformer (SAC-GT), a framework for accurate and reliable indoor localization. SAC-GT integrates a Graph Transformer (GT) model that captures network's spatial topology and signal strength dynamics, with a novel Spatially-Adaptive Conformal Prediction (SACP) method that provides region-specific uncertainty estimates. This allows SAC-GT to produce not only precise two-dimensional (2D) location predictions but also statistically valid confidence regions tailored to varying environmental conditions. Extensive evaluations on a large-scale real-world dataset demonstrate that the proposed SAC-GT solution achieves state-of-the-art localization accuracy while delivering robust and spatially adaptive reliability guarantees.

Spatially-Adaptive Conformal Graph Transformer for Indoor Localization in Wi-Fi Driven Networks

TL;DR

The paper tackles indoor localization with a focus on reliability by introducing SAC-GT, which fuses a Graph Transformer for precise 2D position estimates with Spatially-Adaptive Conformal Prediction to produce region-specific, statistically valid confidence regions. The GT captures spatial topology and RSSI dynamics, while SACP calibrates adaptive radii per region using a held-out calibration set, enabling tailored uncertainty that mirrors environmental heterogeneity. On the SODIndoorLoc benchmark, SAC-GT delivers a median localization error of m and strong 95th-percentile performance, plus an overall predictive coverage of against a target, illustrating robust real-world applicability. The work advances practical indoor localization by providing both high accuracy and reliable uncertainty guarantees, with potential impact on navigation, asset tracking, and safety-critical applications in Wi-Fi driven networks.

Abstract

Indoor localization is a critical enabler for a wide range of location-based services in smart environments, including navigation, asset tracking, and safety-critical applications. Recent graph-based models leverage spatial relationships between Wire-less Fidelity (Wi-Fi) Access Points (APs) and devices, offering finer localization granularity, but fall short in quantifying prediction uncertainty, a key requirement for real-world deployment. In this paper, we propose Spatially-Adaptive Conformal Graph Transformer (SAC-GT), a framework for accurate and reliable indoor localization. SAC-GT integrates a Graph Transformer (GT) model that captures network's spatial topology and signal strength dynamics, with a novel Spatially-Adaptive Conformal Prediction (SACP) method that provides region-specific uncertainty estimates. This allows SAC-GT to produce not only precise two-dimensional (2D) location predictions but also statistically valid confidence regions tailored to varying environmental conditions. Extensive evaluations on a large-scale real-world dataset demonstrate that the proposed SAC-GT solution achieves state-of-the-art localization accuracy while delivering robust and spatially adaptive reliability guarantees.
Paper Structure (11 sections, 6 equations, 6 figures, 2 tables)

This paper contains 11 sections, 6 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. System Model
  3. Problem Formulation
  4. A Spatially-Adaptive Conformal Graph Transformer Model for Indoor Localization
  5. Graph Transformer (GT) Model
  6. Spatially-Adaptive Conformal Prediction (SACP)
  7. Inference Stage
  8. Simulation Results
  9. Dataset
  10. Experimental Settings
  11. Conclusion

Figures (6)

  • Figure 1: System model for indoor localization, showing the graph structure with a mobile user (a user receiver), APs, physical links, and logical links.
  • Figure 2: The proposed Spatially-Adaptive Conformal Graph Transformer (SAC-GT) framework.
  • Figure 3: Localization Error (in meters) versus connectivity threshold $\tau$.
  • Figure 4: Localization error map across regions under the proposed SACP.
  • Figure 5: Effect of conformal significance ($\alpha$) on empirical (test) coverage across regions and globally.
  • ...and 1 more figures