Nitin Jonathan Myers and Robert W. Heath, Jr.
to appear in the Proc. of the IEEE Int. Workshop on Signal Processing Advances in Wireless Communication, Sapporo, Japan, July 3-6,2017.
Initial access at millimeter wave frequencies is a challenging problem due to hardware non-idealities and significant path loss, resulting in low SNR measurements. It is well known that mmWave MIMO channels are sparse in the spatial angle domain and several compressed sensing based algorithms have been proposed for channel estimation exploiting this fact. Most of them, however, ignore hardware impairments like carrier frequency offset, phase noise etc., and fail to perform well when such impairments are considered. In this paper, we develop a compressive channel estimation algorithm for narrowband mmWave systems, that is robust to such non idealities. We address this problem by constructing a tensor that models both the mmWave channel and CFO, and estimate the tensor while still exploiting the sparsity of the mmWave channel. Simulation results show that under the same settings, our method performs better than a recently proposed algorithm that is robust to phase errors.
An implementation of our algorithm can be found here.