Tracking Keystrokes Using Wireless Signals. Chen, B., Yenamandra, V., & Srinivasan, K. In Proceedings of the 13th Annual International Conference on Mobile Systems, Applications, and Services, of MobiSys '15, pages 31-44, 2015. ACM.
Tracking Keystrokes Using Wireless Signals [link]Website  abstract   bibtex   
We implement a passive remote keystroke detection mechanism using only changes in the wireless channel. The detection algorithm does not require the user to wear any active devices nor does it require a change in the user's wireless transmission technique. The receiver system is implemented with five antennas. We cancel the signals received on multiple antennas. The key insight to realizing a fine-grained localization system is to exploit the extremely high sensitivity of cancellation performance (interference cancellation in full-duplex for example) to exact amplitude and phase matching. The receiver design introduces a delay mismatch between the received signal streams to guarantee imperfect cancellation across the transmission bandwidth except at one in-band frequency resulting in a trough in the cancellation spectrum. We detect keystrokes by forming an array of observed trough frequency across different antenna pairs. We implement our receiver system on the NI-based SDR platform. With full-training the receiver detects a keystroke within one key offset with an accuracy of 91.8%. With a short ≈ 10 character training input, this accuracy drops to 85%.
@inProceedings{
 title = {Tracking Keystrokes Using Wireless Signals},
 type = {inProceedings},
 year = {2015},
 identifiers = {[object Object]},
 keywords = {keystrokes,wireless},
 pages = {31-44},
 websites = {http://dx.doi.org/10.1145/2742647.2742673},
 publisher = {ACM},
 city = {New York, NY, USA},
 series = {MobiSys '15},
 id = {6da91093-c94c-3c0d-9cba-134d5dec1946},
 created = {2018-07-12T21:30:57.610Z},
 file_attached = {false},
 profile_id = {f954d000-ce94-3da6-bd26-b983145a920f},
 group_id = {b0b145a3-980e-3ad7-a16f-c93918c606ed},
 last_modified = {2018-07-12T21:30:57.610Z},
 read = {false},
 starred = {false},
 authored = {false},
 confirmed = {true},
 hidden = {false},
 citation_key = {chen:wifikeystrokes},
 source_type = {inproceedings},
 private_publication = {false},
 abstract = {We implement a passive remote keystroke detection mechanism using only changes in the wireless channel. The detection algorithm does not require the user to wear any active devices nor does it require a change in the user's wireless transmission technique. The receiver system is implemented with five antennas. We cancel the signals received on multiple antennas. The key insight to realizing a fine-grained localization system is to exploit the extremely high sensitivity of cancellation performance (interference cancellation in full-duplex for example) to exact amplitude and phase matching. The receiver design introduces a delay mismatch between the received signal streams to guarantee imperfect cancellation across the transmission bandwidth except at one in-band frequency resulting in a trough in the cancellation spectrum. We detect keystrokes by forming an array of observed trough frequency across different antenna pairs. We implement our receiver system on the NI-based SDR platform. With full-training the receiver detects a keystroke within one key offset with an accuracy of 91.8%. With a short ≈ 10 character training input, this accuracy drops to 85%.},
 bibtype = {inProceedings},
 author = {Chen, Bo and Yenamandra, Vivek and Srinivasan, Kannan},
 booktitle = {Proceedings of the 13th Annual International Conference on Mobile Systems, Applications, and Services}
}

Downloads: 0