Optimizing the common-randomness secret key rate in beamformed communication systems

Generating key bits with multiple antenna systems has the potential to dramatically increase the rate at which key bits can be generated. For beamforming systems, this work considers how to probe the channel to maximize the mutual information between channel estimates. For these systems, we provide an optimal probing strategy that uses the eigenvectors of the channel correlation matrix as probing vectors and then optimally allocates the transmission energy to each vector. Since this strategy cannot be implemented in practice, we also propose a practical channel estimation scheme that achieves near-optimal performance. Simulations demonstrate the performance advantage of the proposed strategy relative to that achieved using traditional channel estimation in which the total energy is equally allocated to all probing vectors.