Title:Optimal Signal Processing for Common Randomness Generation over MIMO Gaussian Channels with Applications in Identification

Abstract:Common randomness (CR), as a resource, is not commonly used in existing practical communication systems. In the common randomness framework, both sender and receiver aim to generate a common random variable observable to both, ideally with low error probability. The knowledge of this CR allows us to implement correlated random protocols that can lead to faster and more efficient algorithms. Previous work on CR focused on Single-Input Single-Output (SISO) channels with finite input and output alphabets. As far as we know, no results on continuous and Multiple-Input Multiple-Output (MIMO) channels have yet been established. We characterize CR over Gaussian channels for their practical relevance in many communication situations. We first derive the CR capacity for Gaussian SISO channels. This result is then used for the derivation of the CR capacity in the MIMO case. CR plays a key role in the identification scheme. In contrast to the classical transmission, the decoder is now interested in knowing \emph{whether} a specific message was sent or not, rather than knowing \emph{what} the received message is. In many new applications such as several machine-to-machine and human-to-machine systems and the tactile internet, the identification, also called the post-Shannon scheme, is more efficient than the classical transmission. It has been proved that through CR generation, the post-Shannon communication task allows us to achieve an enormous performance gain that we can take advantage of in several applications.
We consider a correlation-assisted secure identification scheme and develop a lower bound on the corresponding secure identification capacity.