RadEar: A Self-Supervised RF Backscatter System for Voice Eavesdropping and Separation

2026-03-12

AI Summary

This work proposes a passive RF backscatter-based through-wall voice sensing system to address the threat of covert eavesdropping on speech privacy. By designing a dual-resonator passive tag to suppress self-interference and integrating a self-supervised learning framework that operates without ground-truth labels, the system effectively separates and denoises multi-speaker voices. The approach innovatively combines a dual-resonator frequency modulation architecture with a remixing-target-based self-supervised model, enabling high-fidelity voice recovery in real-world scenarios. This integration significantly enhances the practicality and robustness of through-wall voice eavesdropping systems.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Security and Privacy: Privacy-enhancing technologies
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

Eavesdropping on voice conversations presents a growing threat to personal privacy and information security. In this paper, we present RadEar, a novel RF backscatter-based system designed to enable covert voice eavesdropping through walls. RadEar consists of two key components: (i) a batteryless RF backscatter tag covertly deployed inside the target space, and (ii) an RF reader located outside the room that performs signal demodulation, voice separation, and denoising. The tag features a compact, dual-resonator design that achieves energy-efficient frequency modulation for continuous voice eavesdropping while mitigating self-interference by separating excitation and reflection frequencies. To overcome the challenges of weak signal reception and overlapping speech, the RF reader employs self-supervised learning models for voice separation and denoising, trained using a remix-based objective without requiring ground-truth labels. We fabricate and evaluate RadEar in real-world scenarios, demonstrating its ability to recover and separate human speech with high fidelity under practical constraints.

Problem

• voice eavesdropping
• RF backscatter
• speech separation
• privacy threat
• through-wall sensing

Innovation

• RF backscatter
• self-supervised learning
• voice separation
• dual-resonator design
• covert eavesdropping