On channel state feedback for two-hop networks based on low rank matrix recovery

This paper proposes a novel feedback protocol for relay-based two-hop networks, in which the channel state information matrix of the second hop is compressible due to the presence of spatial correlation and distance dependent path loss among the communication channels from the relay nodes to the users. The proposed protocol makes use of recent developments in the fields of low rank matrix recovery and compressed sensing to approximate the channel matrix by a low rank and sparse matrix. As a result, accurate channel state information can be provided to the base station for optimal relay selection, while significantly reducing pilot contamination and feedback overhead. Simulations demonstrate that approximately 50% of the training and feedback overhead can be saved if the compressibility of the channel matrix is taken into account.