127 / 2019-12-15 01:34:00

Millimeter Wave Channel Estimation from Quadratic Measurements

sparse phase retrieval; nonconvex optimization; compressed sensing; MmWave channel estimation

Millimeter wave (mmWave) communication has the potential for satisfying growing demands of air interface capacity and new spectrum allocation of sub-$6$ GHz wireless networks. Recently proposed compressed sensing-based techniques which exploit the inherent sparse scattering nature of mmWave channel. However, these methods work with implicit assumptions on long-term phase coherence that are hard to implement with existing hardware. In this paper, we propose a compressive phase retrieval-based mmWave channel estimation technique which can estimate the sparse mmWave channel parameters from quadratic measurements. Compared with traditional compressed sensing-based channel estimation methods, the proposed method has low-cost hardware implementation and robust to carrier frequency offset caused by high-frequency hardware imperfections. Based on the proposed compressive phase retrieval-based model, we introduce a two-stage algorithm to estimate the mmWave channel parameters (up to the global phase) and then compute the exact solution via the anchor measurements. We provide numerical simulations to demonstrate the efficacy of the approach.