基于WiFi射频信号的侧信道击键识别

WiFi-based side channel keystroke recognition

摘要:

侧信道攻击(side channel attack，SCA)是现代密码学的一个重要分支，在现代信息安全中起到重要作用。随着WiFi技术的日益成熟，研究者尝试在数据传输的同时对周围场景进行感知，实现“感”传一体化。因此WiFi射频信号被用来进行击键识别，其原理是通过不同击键动作引起的WiFi信道状态信息(channel state information, CSI)变化进行键值推理。从无线感知原理、微小动作感知增强2个方面，对基于WiFi的侧信道密码分析进行了梳理介绍；基于微小动作感知增强实现并验证了基于WiFi射频信号的侧信道击键识别系统。实验结果表明，系统在非穿墙模式下对0~9数字按键的识别准确率为85.4%，在穿墙模式下的识别准确率为82.6%。

Abstract:

As one of the important branches of the modern cryptography, side channel attack (SCA) has been playing a significant role in the field of information security. With the maturity of the WiFi technology, researchers attempt to sense the surrounding environment with data transmission at the same time. In recent years, the WiFi radio frequency (RF) signals have been used for keystrokes recognition, the basic idea of which is that the key values can be reasoned out by classifying the channel state information (CSI) patterns produced by the unique gestures and trajectories of users hitting different keys. From the perspectives of wireless perception principle and micro-motion perception amplification, this paper reviewed the WiFi-based side channel cryptoanalysis, and then prototyped and verified a WiFi-based side channel keystroke detection system. Experimental results showed that this system could recognize the keys from 0 to 9 with 85.4% accuracy in the non-wall scenario and with 82.6% accuracy in the through-wall scenario, respectively.