Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FM-RME: Foundation Model Empowered Radio Map Estimation

Dong Yang, Yue Wang, Songyang Zhang, Yingshu Li, Zhipeng Cai, Zhi Tian

TL;DR

A new foundation model, characterized by self-supervised pre-training on diverse data for zero-shot generalization, enabling multi-dimensional radio map estimation (FM-RME), which exhibits desired learning performance across diverse datasets and zero-shot generalization capabilities beyond existing RME methods.

Abstract

Traditional radio map estimation (RME) techniques fail to capture multi-dimensional and dynamic characteristics of complex spectrum environments. Recent data-driven methods achieve accurate RME in spatial domain, but ignore physical prior knowledge of radio propagation, limiting data efficiency especially in multi-dimensional scenarios. To overcome such limitations, we propose a new foundation model, characterized by self-supervised pre-training on diverse data for zero-shot generalization, enabling multi-dimensional radio map estimation (FM-RME). Specifically, FM-RME builds an effective synergy of two core components: a geometry-aware feature extraction module that encodes physical propagation symmetries, i.e., translation and rotation invariance, as inductive bias, and an attention-based neural network that learns long-range correlations across the spatial-temporal-spectral domains. A masked self-supervised multi-dimensional pre-training strategy is further developed to learn generalizable spectrum representations across diverse wireless environments. Once pre-trained, FM-RME supports zero-shot inference for multi-dimensional RME, including spatial, temporal, and spectral estimation, without scenario-specific retraining. Simulation results verify that FM-RME exhibits desired learning performance across diverse datasets and zero-shot generalization capabilities beyond existing RME methods.

FM-RME: Foundation Model Empowered Radio Map Estimation

TL;DR

A new foundation model, characterized by self-supervised pre-training on diverse data for zero-shot generalization, enabling multi-dimensional radio map estimation (FM-RME), which exhibits desired learning performance across diverse datasets and zero-shot generalization capabilities beyond existing RME methods.

Abstract

Traditional radio map estimation (RME) techniques fail to capture multi-dimensional and dynamic characteristics of complex spectrum environments. Recent data-driven methods achieve accurate RME in spatial domain, but ignore physical prior knowledge of radio propagation, limiting data efficiency especially in multi-dimensional scenarios. To overcome such limitations, we propose a new foundation model, characterized by self-supervised pre-training on diverse data for zero-shot generalization, enabling multi-dimensional radio map estimation (FM-RME). Specifically, FM-RME builds an effective synergy of two core components: a geometry-aware feature extraction module that encodes physical propagation symmetries, i.e., translation and rotation invariance, as inductive bias, and an attention-based neural network that learns long-range correlations across the spatial-temporal-spectral domains. A masked self-supervised multi-dimensional pre-training strategy is further developed to learn generalizable spectrum representations across diverse wireless environments. Once pre-trained, FM-RME supports zero-shot inference for multi-dimensional RME, including spatial, temporal, and spectral estimation, without scenario-specific retraining. Simulation results verify that FM-RME exhibits desired learning performance across diverse datasets and zero-shot generalization capabilities beyond existing RME methods.
Paper Structure (29 sections, 15 equations, 4 figures, 1 table)

This paper contains 29 sections, 15 equations, 4 figures, 1 table.

Table of Contents

  1. Introduction
  2. Signal Model and Problem Formulation
  3. Spatial-Temporal-Spectral Radio Map
  4. Graph Representation of Radio Map
  5. Problem Formulation
  6. FM-RME Architecture
  7. Self-Supervised Pre-Training (SSPT)
  8. Spatial Domain Masked Estimation
  9. Temporal Domain Masked Estimation
  10. Spectral Domain Masked Estimation
  11. Geometry-Aware Feature Extraction (GAFE)
  12. Message Passing
  13. Scalar Message
  14. Vector Message
  15. Aggregation and Node Update
  16. ...and 14 more sections

Figures (4)

  • Figure 1: Multi-dimensional radio map estimation: RME aims to reconstruct full PSD radio maps from sparse measurements, enabling RME tasks cross spatial-temporal-spectral domains.
  • Figure 2: The overview of FM-RME framework.
  • Figure 3: RMSE performance comparison for spatial estimation and zero-shot generalization.
  • Figure 4: RMSE comparison for temporal and spectral estimation.