Anti-Spoofing Voice Commands: A Generic Wireless Assisted Design
summary
Prototype system builds three WISE versions using RFID, Wi‑Fi, and inaudible acoustic sensing and adds one‑time random noise to shelter transceivers. RFID uses noise‑like CW to hide EPC, Wi‑Fi rotates constellation per packet and skips preamble, acoustic uses OFDM with similar scrambling. A neural net fuses audio plus sensing to verify liveness and identity. Tested in six real environments with 12 users, verification stays >90% and dual‑channel replay is rejected >99%.
AI-written summary · about
chapters (5)
related talks
30mHidden Voice Commands
39m Put Words In My Mouth
45mAnalogy Cyber Security—From 0101 to Mixed Signals
24mInaudible Voice Commands: The Long-Range Attack and Defense
24mCommanderSong: A Systematic Approach for Practical Adversarial Voice Recognition
33mYour voice confirms my identity
42mYour Wi-Fi Is the Eavesdropper’s Radar: How to Counter Privacy Threats...
06mUsing Sonar for Liveness Detection to Protect Smart Speakers against Remote Attackers
