One Walk is All You Need: Data-Efficient 3D RF Scene Reconstruction with Human Movements

2025-11-21

AI Summary

This work addresses the key bottlenecks in radio frequency (RF) scene reconstruction—namely, reliance on dense static sampling and poor robustness to occlusions. We propose a novel paradigm that achieves high-fidelity 3D static radiance field reconstruction from only a single, brief (60-second) human walk-through. Crucially, we treat human motion not as noise but as informative signal, introducing a composite 3D Gaussian splatting decomposition framework that explicitly models and separates dynamic human effects from static scene geometry. Our method jointly learns dynamic and static factors directly from raw RF time-series measurements, enabling end-to-end reconstruction without any static reference scans. Experiments demonstrate an SSIM of 0.96—12% higher than state-of-the-art dense-sampling methods—while drastically reducing acquisition cost and overcoming occlusion limitations.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Learning with Manifolds

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Reconstructing 3D Radiance Field (RF) scenes through opaque obstacles is a long-standing goal, yet it is fundamentally constrained by a laborious data acquisition process requiring thousands of static measurements, which treats human motion as noise to be filtered. This work introduces a new paradigm with a core objective: to perform fast, data-efficient, and high-fidelity RF reconstruction of occluded 3D static scenes, using only a single, brief human walk. We argue that this unstructured motion is not noise, but is in fact an information-rich signal available for reconstruction. To achieve this, we design a factorization framework based on composite 3D Gaussian Splatting (3DGS) that learns to model the dynamic effects of human motion from the persistent static scene geometry within a raw RF stream. Trained on just a single 60-second casual walk, our model reconstructs the full static scene with a Structural Similarity Index (SSIM) of 0.96, remarkably outperforming heavily-sampled state-of-the-art (SOTA) by 12%. By transforming the human movements into its valuable signals, our method eliminates the data acquisition bottleneck and paves the way for on-the-fly 3D RF mapping of unseen environments.

Problem

• Reconstructing 3D scenes through obstacles with minimal data collection
• Using human motion as signal instead of noise for reconstruction
• Eliminating need for thousands of static measurements with single walk

Innovation

• Uses human walk as signal for reconstruction
• Employs composite 3D Gaussian Splatting factorization
• Learns static geometry from dynamic RF streams