A Survey of User Authentication Based on Channel State Information

Table 4

Stillness-based user authentication applications.

System	Users	Signal/preprocessing	Experimental scene	Purpose/classification	Performance
Liu et al. [84]	12 locations	Amplitude/temporal correlation analysis	A laboratory () and an apartment ()	Identity recognition stranger identification/SVM	Attack detection: 92%; Authentication accuracy: 98.4%
WiPIN [56]	30	Amplitude/Butterworth filter	A room ()	Identity recognition/SVM	Two people: 100%; 30 people: 92%
Abyaneh et al. [57]	11 locations	Amplitude and phase	An apartment, a garage, and the ramp	Identity recognition/LocNet	Apartment: 85.35%; Garage: 98.5%
BodyPIN [85]	30	Amplitude and phase/Butterworth low-pass filter	A room ()	Identity recognition/SVM	30 users: 92%
Liu et al. [61]	20	Amplitude/EMD-based filter	A university office ()	Identity recognition, stranger identification/KNN	20 people: 93%; Random attacks: 92.14% with 5% FP; Imitation attacks: 89.24% with 5% FP
Wang et al. [86]	10	Amplitude/PCA	A laboratory ()	Identity recognition/softmax regression	10 people: 97.5% (respiration); 90.4% (gait)
CP-ID [66]	6	Phase/PCA	A typical office ()	Identity recognition/SVM	From 2 to 5 people: 84% to 65%
BioID [68]	5	Amplitude/Butterworth low-pass filter, PCA, and DWT	An office	Identity recognition/KNN	5 people: 90%