Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Graph-Based Floor Separation Using Node Embeddings and Clustering of WiFi Trajectories

Rabia Yasa Kostas, Kahraman Kostas

TL;DR

This work tackles blind floor separation in GPS-denied multistory buildings using Wi-Fi fingerprint trajectories. It introduces a fully data-driven graph framework that builds a trajectory graph, learns Node2Vec embeddings, and clusters with automatic floor-number estimation, without relying on building metadata. The approach achieves state-of-the-art performance across Huawei and UJIIndoorLoc datasets, with the WBDE distance estimator enhancing robustness and, in the UJI setting, achieving results comparable to geometry-based graphs. The findings demonstrate that topology-driven, embedding-based clustering can reliably uncover floor structure from noisy RSSI data, offering a scalable solution for indoor vertical localization.

Abstract

Vertical localization, particularly floor separation, remains a major challenge in indoor positioning systems operating in GPS-denied multistory environments. This paper proposes a fully data-driven, graph-based framework for blind floor separation using only Wi-Fi fingerprint trajectories, without requiring prior building information or knowledge of the number of floors. In the proposed method, Wi-Fi fingerprints are represented as nodes in a trajectory graph, where edges capture both signal similarity and sequential movement context. Structural node embeddings are learned via Node2Vec, and floor-level partitions are obtained using K-Means clustering with automatic cluster number estimation. The framework is evaluated on multiple publicly available datasets, including a newly released Huawei University Challenge 2021 dataset and a restructured version of the UJIIndoorLoc benchmark. Experimental results demonstrate that the proposed approach effectively captures the intrinsic vertical structure of multistory buildings using only received signal strength data. By eliminating dependence on building-specific metadata, the proposed method provides a scalable and practical solution for vertical localization in indoor environments.

Graph-Based Floor Separation Using Node Embeddings and Clustering of WiFi Trajectories

TL;DR

This work tackles blind floor separation in GPS-denied multistory buildings using Wi-Fi fingerprint trajectories. It introduces a fully data-driven graph framework that builds a trajectory graph, learns Node2Vec embeddings, and clusters with automatic floor-number estimation, without relying on building metadata. The approach achieves state-of-the-art performance across Huawei and UJIIndoorLoc datasets, with the WBDE distance estimator enhancing robustness and, in the UJI setting, achieving results comparable to geometry-based graphs. The findings demonstrate that topology-driven, embedding-based clustering can reliably uncover floor structure from noisy RSSI data, offering a scalable solution for indoor vertical localization.

Abstract

Vertical localization, particularly floor separation, remains a major challenge in indoor positioning systems operating in GPS-denied multistory environments. This paper proposes a fully data-driven, graph-based framework for blind floor separation using only Wi-Fi fingerprint trajectories, without requiring prior building information or knowledge of the number of floors. In the proposed method, Wi-Fi fingerprints are represented as nodes in a trajectory graph, where edges capture both signal similarity and sequential movement context. Structural node embeddings are learned via Node2Vec, and floor-level partitions are obtained using K-Means clustering with automatic cluster number estimation. The framework is evaluated on multiple publicly available datasets, including a newly released Huawei University Challenge 2021 dataset and a restructured version of the UJIIndoorLoc benchmark. Experimental results demonstrate that the proposed approach effectively captures the intrinsic vertical structure of multistory buildings using only received signal strength data. By eliminating dependence on building-specific metadata, the proposed method provides a scalable and practical solution for vertical localization in indoor environments.
Paper Structure (40 sections, 3 equations, 4 figures, 3 tables)

This paper contains 40 sections, 3 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Barometric Pressure, IMU, and Multimodal Approaches
  4. Wi-Fi RSSI-Based Floor Detection
  5. Graph-Based Spatial Modeling and Clustering
  6. Graph Embeddings for Wi-Fi-Based Floor Estimation
  7. METHODOLOGY
  8. Problem Formulation
  9. Datasets
  10. Huawei University Challenge 2021
  11. Validation on UJIIndoorLoc Dataset
  12. Edge Weighting and Scenario Design
  13. Motivation: Role of Distance Estimation in Graph Construction
  14. Clustering and Community Detection Methods
  15. Baseline Community Detection Algorithms
  16. ...and 25 more sections

Figures (4)

  • Figure 1: Overview of the experimental workflow. The trajectory graph is used as input for both the baseline community detection algorithms (Louvain, Leiden, Infomap, Label Propagation, Fast Greedy) and the proposed Node2Vec+KMeans pipeline. While baseline methods operate directly on the graph structure, the proposed approach first transforms the graph into a 32-dimensional embedding space using Node2Vec, followed by clustering with K-Means.
  • Figure 2: Mapped F1 scores of different community detection and representation learning algorithms evaluated on the Huawei dataset. Error bars indicate 95% confidence intervals estimated from 1,000 bootstrap iterations.
  • Figure 3: Mapped F1 scores of different community detection and representation learning algorithms evaluated on the UJI dataset. Error bars indicate 95% confidence intervals estimated from 1,000 bootstrap iterations.
  • Figure 4: Confusion matrices for the six clustering methods. Strong diagonal dominance in the Node2Vec matrix contrasts with the dispersed predictions of the baselines.