DoRF: Doppler Radiance Fields for Robust Human Activity Recognition Using Wi-Fi

2025-07-16

AI Summary

Weak generalization capability of Wi-Fi CSI-based human activity recognition (HAR)—particularly across unseen environments and users—remains a critical challenge. To address this, we propose the Doppler Radiance Field (DoRF), the first method to elevate one-dimensional Doppler velocity projections into a three-dimensional implicit motion representation, inspired by neural radiance fields (NeRF) to construct a spatially structured, continuous volumetric representation. DoRF jointly learns CSI preprocessing, Doppler spectrum extraction, and implicit neural representation in an end-to-end manner, eliminating the need for explicit pose estimation or handcrafted features. Extensive experiments demonstrate that DoRF significantly improves recognition accuracy and robustness under cross-environment and cross-user settings, achieving an average 9.2% gain over state-of-the-art baselines. These results validate DoRF’s strong generalization capacity and practical deployability for real-world Wi-Fi sensing applications.

Technology Category

• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Computer Vision: Motion & Tracking
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Wi-Fi Channel State Information (CSI) has gained increasing interest for remote sensing applications. Recent studies show that Doppler velocity projections extracted from CSI can enable human activity recognition (HAR) that is robust to environmental changes and generalizes to new users. However, despite these advances, generalizability still remains insufficient for practical deployment. Inspired by neural radiance fields (NeRF), which learn a volumetric representation of a 3D scene from 2D images, this work proposes a novel approach to reconstruct an informative 3D latent motion representation from one-dimensional Doppler velocity projections extracted from Wi-Fi CSI. The resulting latent representation is then used to construct a uniform Doppler radiance field (DoRF) of the motion, providing a comprehensive view of the performed activity and improving the robustness to environmental variability. The results show that the proposed approach noticeably enhances the generalization accuracy of Wi-Fi-based HAR, highlighting the strong potential of DoRFs for practical sensing applications.

Problem

• Improve Wi-Fi-based human activity recognition robustness
• Reconstruct 3D motion from 1D Doppler velocity projections
• Enhance generalization accuracy for practical sensing applications

Innovation

• Reconstruct 3D latent motion from Doppler projections
• Use neural radiance fields for Wi-Fi sensing
• Create uniform Doppler radiance field (DoRF)