EarFisher: Detecting Wireless Eavesdroppers by Stimulating and Sensing Memory EMR
summary
System induces bait web-page bursts and uses a BladeRF-based detector to sense memory EMR peaks produced when a receiver writes every packet to RAM, separating devices by folding spectra into (f0, fi) subclock features. Detection compares stimulus and baseline FFT windows and flags devices whose memory EMR rises under stimulus; evaluation in a 1,600 ft² testbed finds 8/9 DDR3 and 7/9 DDR4 eavesdroppers detected with recall >93% and no false alarms, and the design resists normal workloads and coexisting devices.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:37Eavesdropping Is a Fundamental Threat to Wireless Security
- 1:37State-of-The-Art Detectors
- 2:15Key Insight
- 3:23The EarFisher System
- 4:01Key Challenges
- 5:52Sensing Memory EMR
- 6:36Separating Memory EMR • Exploit clock fingerprints to separate EMRS
- 9:12Inferring Eavesdropper's Response
- 10:21Experiment Results
- 11:33Conclusion
related talks
42mYour Wi-Fi Is the Eavesdropper’s Radar: How to Counter Privacy Threats...
55mReliable and Secure Embedded Systems
06mAnti-Spoofing Voice Commands: A Generic Wireless Assisted Design
33mFor Smarter Authentication, We Might Need to Use ...
55mEtherify, sonify: soft tempest at home and on the farm
16mEchoLLM: LLM-Augmented Acoustic Eavesdropping Attack on Bone Conduction...
18mdetecting and defending against state actor surveillance
24mEnhancing Physical Perimeter Defense Using SDR