Exploiting Common Sparsity for Frequency-Domain Wideband Channel Estimation at mmWave.

Hybrid architectures reduce power consumption for MIMO systems at millimeter wave (mmWave) with respect to all-digital solutions. Fast configuration of hybrid arrays is one of the main challenges to be solved to unlock the potential of mmWave in high mobility scenarios. Channel estimation is one of the means of obtaining the information needed to configure these arrays. Prior work on channel estimation at mmWave considers a frequency-flat narrowband channel model, although the the mmWave channel is intrinsically wideband and frequency-selective. In this paper, we design a frequency-domain algorithm that leverages the sparsity in the channel to estimate the frequency-selective mmWave channel. We evaluate this technique by simulation and study its complexity, comparing the trade-offs complexity-overhead-performance to those of previous approaches. We conclude that the proposed approach outperforms previous algorithms, providing a better performance at a lower computational cost.