Millimeter wave link configuration in practical scenarios




Xie, Hongxiang, Ph. D.

Journal Title

Journal ISSN

Volume Title



Acquiring channel state information (CSI) for link configuration in wideband millimeter wave (mmWave) massive multiple-input-multiple-output (MIMO) systems with hybrid architectures is challenging, due to the high dimensions of the channel matrices, the low signal-to-noise ratio (SNR) before beamforming, the various hardware constraints and the high mobility in the vehicular context. Previous work in this area exploits channel sparsity, statistical priors or side information to reduce the overhead associated to initial channel estimation or channel tracking. These works consider, however, a system model that neglects hardware imperfections. In addition, many of the proposed solutions are unable to operate in some realistic scenarios, such as vehicle-to-everything (V2X) communications. In this dissertation, we develop new signal processing solutions that can enable low-overhead mmWave link configuration under various disturbances and practical limitations, e.g., hardware impairments, calibration errors, beam squint effect, channel blockage, high mobility, to name a few.

In the first part of this dissertation, we focus on the problem of wideband channel estimation for mmWave MIMO systems with different hardware imperfections. We first design a dictionary learning aided channel estimation strategy for wideband mmWave MIMO systems by explicitly considering the hardware uncertainties and calibration errors, and then derive algorithms that learn the optimal sparsifying dictionaries for channel representation and estimation. In a second contribution of this part, we further develop a dictionary learning aided compressive channel estimation scheme for mmWave MIMO systems by incorporating beam squint into the model of array responses. Numerical results show the proposed solutions can adapt to the practical scenarios and help reduce the overhead associated with channel estimation significantly.

In the second part of this dissertation, we deal with the problem of wideband channel tracking for mmWave MIMO systems with or without the impact of blockage. We first introduce statistical channel models that include the evolution models for channel gains and angles of arrival/departure, as well as the statistics of blockage events. Then, we design novel blockage detection schemes and efficient Bayesian channel tracking algorithms to facilitate the low-overhead tracking with or without blockage. Numerical results corroborate that the proposed solutions achieve better channel tracking performance even in mobile scenarios that suffer from highly dynamic blockage events.


LCSH Subject Headings