Applying Beamforming to Address Temporal Correlation in Wireless Channel Characterization-Based Secret Key Generation

Wireless secret key generation (WKG) methods based on communications channel characterizations have gained significant interest as a mechanism for providing secure point-to-point communications. Ultra-wide band (UWB) WKG solutions provide an advantage over narrow- and wide-band approaches in that they can generate secure keys even when Alice and Bob, the two communicating parties, and all objects within the environment in which they exist, are stationary (i.e., in scenarios involving zero movement). This creates a secondary problem in that the high temporal correlations of such environments can lead to successive WKG processes producing identical or nearly identical secret keys, whereas the ideal is that all generated keys from successive WKG processes should be independent. This work shows that one method available to address this problem is to move to multiple-input multiple-output (MIMO)-based antenna systems and the use of random beamforming. The security of the resulting approach is assessed with respect to the worst case scenario whereby the eavesdropper, Eve, is assumed to possess perfect knowledge of Alice and Bob's MIMO channel and is able to estimate, with some degree of error, the coefficients of Alice and Bob's independent random beamformers.