연구 분야: Analysis
학회: 2025 23rd Mediterranean Communication and Computer Networking Conference (MedComNet)
Exploiting Channel State Information (CSI) from ubiquitous Wi-Fi traffic facilitates widespread environmental sensing, possibly leading to unauthorized surveillance and severe privacy concerns. A promising countermeasure to this issue is to obfuscate the CSI by transmitting distorted versions of the signals that block malicious sensing operations without affecting the communication. However, this also hinders legitimate sensing. This work demonstrates the feasibility of CSI de-obfuscation, where authorized receivers can compensate for pre-distortion to recover sensing capabilities. As a case study, we analyze device-free localization with two CSI obfuscation patterns and three levels of distortion strength, verifying that stronger distortion is more effective but harder to reverse. In general, proper de-obfuscation can almost completely restore localization accuracy; however, it becomes more challenging when the obfuscation pattern introduces notches in the CSI, i.e., when some subcarriers get close to zero. Early results show the practical feasibility of de-obfuscation strategies, providing valuable insight into the development of obfuscation patterns and contributing to the design of secure sensing systems.
| 발행 연도 | 2025년 |
|---|---|
| 인용수 | 21 |
| 출판 국가 | Italy |
| 사이트 | IEEE |
| 좋아요 수 | 0 |