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

2025-11-25

AI Summary

In indoor radio map (RM) construction, electromagnetic simulations suffer from high computational latency, while existing learning-based methods rely on sparse measurements and homogeneous material assumptions, failing to capture multipath propagation and structural heterogeneity. To address this, we propose a sampling-free physics-informed diffusion model. Our method incorporates multipath physical priors—including diffraction points, highly transmissive boundaries, and line-of-sight contours—and encodes access point positions and material reflection/transmission coefficients as conditional prompts. A boundary-weighted objective function guides the generative process to achieve high-fidelity modeling of non-stationary field discontinuities. Experiments demonstrate significant improvements in RM accuracy and received signal strength (RSS)-based localization performance across diverse real-world layouts and material configurations. Moreover, the model exhibits strong cross-scenario generalization, achieving, for the first time, high-accuracy radio map generation without requiring any on-site measurements.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Intelligent Robots: Localization, Mapping, and Navigation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

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.

Problem

• Constructs indoor radio maps using physics-informed diffusion models
• Overcomes limitations of EM solvers and learning-based methods
• Models heterogeneous environments with multipath-critical priors

Innovation

• Physics-informed diffusion model for radio map construction
• Conditions on AP positions and material coefficients
• Incorporates multipath-critical priors via conditional channels