Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

iRadioDiff: Physics-Informed Diffusion Model for Indoor Radio Map Construction and Localization

Xiucheng Wang, Tingwei Yuan, Yang Cao, Nan Cheng, Ruijin Sun, Weihua Zhuang

TL;DR

iRadioDiff introduces a sampling-free indoor RM construction method by embedding physics of EM propagation into a diffusion model conditioned on AP positions and material coefficients. It extracts multipath-critical priors (diffraction points, strong transmission boundaries, and LoS contours) to guide generation through conditioning channels and boundary-aware losses, enabling accurate, nonstationary RM reconstructions. Empirical results show state-of-the-art RM construction and improved RSSI-based localization with robust generalization across layouts and materials, while ablations confirm the value of the physics priors. The approach promises scalable, real-time RM generation for environment-aware localization in complex indoor spaces, with planned extensions to multi-AP, 3D, and dynamic scenarios.

Abstract

Radio maps (RMs) serve as environment-aware electromagnetic (EM) representations that connect scenario geometry and material properties to the spatial distribution of signal strength, enabling localization without costly in-situ measurements. However, constructing high-fidelity indoor RMs remains challenging due to the prohibitive latency of EM solvers and the limitations of learning-based methods, which often rely on sparse measurements or assumptions of homogeneous material, which are misaligned with the heterogeneous and multipath-rich nature of indoor environments. To overcome these challenges, we propose iRadioDiff, a sampling-free diffusion-based framework for indoor RM construction. iRadioDiff is conditioned on access point (AP) positions, and physics-informed prompt encoded by material reflection and transmission coefficients. It further incorporates multipath-critical priors, including diffraction points, strong transmission boundaries, and line-of-sight (LoS) contours, to guide the generative process via conditional channels and boundary-weighted objectives. This design enables accurate modeling of nonstationary field discontinuities and efficient construction of physically consistent RMs. Experiments demonstrate that iRadioDiff achieves state-of-the-art performance in indoor RM construction and received signal strength based indoor localization, which offers effective generalization across layouts and material configurations. Code is available at https://github.com/UNIC-Lab/iRadioDiff.

iRadioDiff: Physics-Informed Diffusion Model for Indoor Radio Map Construction and Localization

TL;DR

iRadioDiff introduces a sampling-free indoor RM construction method by embedding physics of EM propagation into a diffusion model conditioned on AP positions and material coefficients. It extracts multipath-critical priors (diffraction points, strong transmission boundaries, and LoS contours) to guide generation through conditioning channels and boundary-aware losses, enabling accurate, nonstationary RM reconstructions. Empirical results show state-of-the-art RM construction and improved RSSI-based localization with robust generalization across layouts and materials, while ablations confirm the value of the physics priors. The approach promises scalable, real-time RM generation for environment-aware localization in complex indoor spaces, with planned extensions to multi-AP, 3D, and dynamic scenarios.

Abstract

Radio maps (RMs) serve as environment-aware electromagnetic (EM) representations that connect scenario geometry and material properties to the spatial distribution of signal strength, enabling localization without costly in-situ measurements. However, constructing high-fidelity indoor RMs remains challenging due to the prohibitive latency of EM solvers and the limitations of learning-based methods, which often rely on sparse measurements or assumptions of homogeneous material, which are misaligned with the heterogeneous and multipath-rich nature of indoor environments. To overcome these challenges, we propose iRadioDiff, a sampling-free diffusion-based framework for indoor RM construction. iRadioDiff is conditioned on access point (AP) positions, and physics-informed prompt encoded by material reflection and transmission coefficients. It further incorporates multipath-critical priors, including diffraction points, strong transmission boundaries, and line-of-sight (LoS) contours, to guide the generative process via conditional channels and boundary-weighted objectives. This design enables accurate modeling of nonstationary field discontinuities and efficient construction of physically consistent RMs. Experiments demonstrate that iRadioDiff achieves state-of-the-art performance in indoor RM construction and received signal strength based indoor localization, which offers effective generalization across layouts and material configurations. Code is available at https://github.com/UNIC-Lab/iRadioDiff.

Paper Structure

This paper contains 12 sections, 3 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Preliminary and System Model
  3. Diffusion Model
  4. System Model and Problem Formulation
  5. Physics Informed DM for Indoor RM Generation
  6. Physics-Informed Environment Feature Extraction
  7. Physics Enhanced Diffusion Model
  8. Experiment Results
  9. Datasets and Implementation Details
  10. Performance Comparison
  11. Ablation Study
  12. Conclusion

Figures (4)

  • Figure 1: Illustration of the multipath boundary extraction. The darker boundary means a higher reflectance and transmittance factor for both figures of reflectance and transmittance.
  • Figure 2: Illustration of the iRadioDiff framework.
  • Figure 3: Illustration of the generated RM from different methods in antenna generalization scenarios.
  • Figure 4: Illustration of the generated RM from different methods in zero-shot layout generalization scenarios.