Kronecker Compressive Sensing for OFDM Channel Estimation in Millimeter Wave Channels

Abstract

We develop a channel estimation approach utilizing the joint sparsity in the virtual channel for OFDM channel estimation in a frequency selective millimeter wave channel. Our approach addresses possible non-uniformly spaced base station antenna arrays by representing them as a sampling of a larger virtual antenna array with uniform spacing. A virtual channel representation is developed associated with the virtual array where the channel is presumed sparse. A Kronecker Compressive Sensing approach is adopted to estimate the virtual channel represented by receive angle of arrivals and delays. This approach is shown to out perform the least squares channel estimate and newly proposed single antenna channel estimators implemented on a per antenna basis.