Enhancing Wireless Secret-Key Generation Through Time-Frequency Analysis Using Wavelet Coherence

The reciprocity nature of a wireless channel has been leveraged to generate and establish secret keys between two devices communicating over the channel. This is typically done by having the two devices exchange probing signals to measure their channel state information (CSI), which each then uses to derive secret keys. While extensive research on secret key generation primarily focuses on theoretical and simulation analyses, the practical implementation and adoption of key generation encounter obstacles arising from hardware constraints and varying channel conditions. To address this research gap, we experimentally investigate the channel measurements by collecting CSI data using resource-constrained devices. Our experiments reveal a significant degradation in the correlation between the collected CSI from two low-cost devices limiting the key generation performance. Through our experimental investigations, we first demonstrate that wavelet coherence provides insights on the reciprocity of the channel measurements on both time and frequency. We then propose a new wavelet coherence-based CSI data reconstruction technique that uses wavelet coherence and time-lagged cross-correlation to reconstruct CSI data that is consistent between the two participating devices, resulting in significant improvement in the channel measurements. Additionally, we propose a secret-key generation scheme that exploits the proposed wavelet coherence-based CSI data reconstruction, yielding a significant reduction in the bit error rates and an increase in the key generation rates.