RaKey: A Millimeter-Wave Radar-Based Keystroke User Authentication
Junxian Ma, Zhenyu Liu, He Yuan Huang, Kangzheng Chen · IEEE Internet of Things Journal · 2025
Keystroke authentication is essential in daily IoT applications. However, two existing challenges are the vulnerability in dynamic environments and the poor performance with limited data. RaKey, a novel millimeter-wave radar (mmWave radar) keystroke authentication scheme, is proposed as a secondary encryption to address these issues. The mmWave radar, with the advantages of being contactless, privacy-preserving and light-insensitive, can authenticate robustly in dynamic scenarios. First, the mmWave radar signal model of keystroke behavior is presented and analyzed for the first time. Second, the micro-Doppler Maps of Overall keystroke behavior (ODMs) and Finger keystroke rhythm (FDMs) are introduced to enhance the expression of keystroke identity features. And RK-AW, an FDMs extraction method based on wavelet with adaptive selection of decomposition level, is proposed to separate the finger keystroke rhythm from overall behavior. Third, to solve the challenge of authentication performance with limited data, RK-Net, a multimodal few-shot network for radar-based keystroke that incorporates complementation signatures and an MCA-LSTM block to fuse multimodal features and thereby enhance the capability of feature extraction, is proposed. Extensive experiments demonstrate that RaKey can achieve considerable authentication accuracy in verifying identities with limited data, while maintaining robustness against environmental variations and potential attacks.