摘要
针对基于信道状态信息(Channel State Information,CSI)的击键行为识别易被环境及噪声干扰,导致击键波形特征的时间稳定性与空间独特性受到损害的问题,提出一种基于空间特征的增强型击键行为识别方法。该方法首先测定击键信号的波动随设备内在及外在环境噪声的非线性增长关系,确定有效击键波形;其次,根据不同键位的空间感知能力,建立一种更稳定、更独特的击键特征刻画方法;最后,充分利用多根天线,构建基于K最邻近(K Nearest Neighbor,KNN)算法的击键行为识别模型,获得最终识别结果。实验结果表明,所提方法的识别准确率能够达到92%以上,相比传统的击键检测方法提高5%~8%,能够有效地提升识别的稳定性与准确性。
Aiming at the problem that keystroke behavior recognition based on CSI is easily disturbed by the environment and noise,which damages the time stability and spatial uniqueness of keystroke waveform characteristic,an enhanced keystroke behavior recognition method based on the spatial characteristic is proposed.The nonlinear growth relationship between the fluctuation of keystroke signal and the internal and external environmental noise of the equipment is firstly measured,and then the effective keystroke waveform is ensured;Secondly,according to the spatial perception ability of different keys,a more stable and unique keystroke characteristic characterization method is established;Finally,the keystroke behavior recognition model based on KNN is constructed to obtain the recognition result by making use of multiple antennas.Experimental results show that the recognition accuracy of the method can reach more than 92%,which is 5%-8%higher than the traditional keystroke recognition method.Therefore,this method can effectively improve the stability and accuracy of the recognition.
作者
陶志勇
张洁
刘影
张旭
TAO Zhiyong;ZhANG Jie;LIU Ying;ZHANG Xu(School of Electronic and Information Engineering,Liaoning Technical University,Huludao Liaoning 125105,China;College of Network and Communication Engineering,Jinling Institute of Technology,Nanjing Jiangsu 211169,China)
出处
《传感技术学报》
CAS
CSCD
北大核心
2022年第12期1678-1685,共8页
Chinese Journal of Sensors and Actuators
基金
辽宁省教育厅项目(LJKZ0349)。
关键词
击键识别
信道状态信息
空间特征
时间稳定性
空间独特性
keystroke recognition
channel state information
spatial characteristic
time stability
spatial uniqueness
