Technology Category
Natural Language Processing: GenerationMachine Learning: Large Multimodal Models (LMMs)Intelligent Robots: Multimodal Perception & Sensor Fusion
One Snapshot is All You Need: A Generalized Method for mmWave Signal Generation
🤖 AI Summary
To address the lack of high-quality, full-resolution raw signal datasets in millimeter-wave (mmWave) wireless sensing, this paper proposes mmGen—a novel framework that, for the first time, enables end-to-end, physics-based synthesis of high-fidelity FMCW raw signals directly from a single-frame 3D mesh scene containing both human subjects and environmental geometry. mmGen integrates material electromagnetic properties, antenna radiation patterns, and realistic multipath propagation and reflection modeling, ensuring physical consistency and cross-scene generalizability. Evaluated across three real-world environments, mmGen-synthesized signals achieve Range-Angle heatmap similarity >0.91 and micro-Doppler spectrogram similarity >0.89—substantially outperforming existing point-cloud– or RA-map–based generative approaches. As the first scalable, physically grounded, and fully annotated raw-signal generation paradigm for mmWave sensing, mmGen establishes a new standard for data synthesis in this domain.
Technology Category
Application Category
📝 Abstract
Wireless sensing systems, particularly those using mmWave technology, offer distinct advantages over traditional vision-based approaches, such as enhanced privacy and effectiveness in poor lighting conditions. These systems, leveraging FMCW signals, have shown success in human-centric applications like localization, gesture recognition, and so on. However, comprehensive mmWave datasets for diverse applications are scarce, often constrained by pre-processed signatures (e.g., point clouds or RA heatmaps) and inconsistent annotation formats. To overcome these limitations, we propose mmGen, a novel and generalized framework tailored for full-scene mmWave signal generation. By constructing physical signal transmission models, mmGen synthesizes human-reflected and environment-reflected mmWave signals from the constructed 3D meshes. Additionally, we incorporate methods to account for material properties, antenna gains, and multipath reflections, enhancing the realism of the synthesized signals. We conduct extensive experiments using a prototype system with commercial mmWave devices and Kinect sensors. The results show that the average similarity of Range-Angle and micro-Doppler signatures between the synthesized and real-captured signals across three different environments exceeds 0.91 and 0.89, respectively, demonstrating the effectiveness and practical applicability of mmGen.
Problem
Research questions and friction points this paper is trying to address.
Generates realistic mmWave signals from 3D meshes
Addresses scarcity of diverse mmWave datasets
Improves signal realism with material and multipath modeling
Innovation
Methods, ideas, or system contributions that make the work stand out.
Generalized framework for mmWave signal generation
Synthesizes signals from 3D meshes with realism
Incorporates material properties and multipath reflections
🔎 Similar Papers
No similar papers found.
