Channel Estimation and Hybrid Precoding for Millimeter Wave Cellular Systems

Millimeter Wave (mmWave) cellular will enable gigabit-per-second data rates thanks to the large bandwidth potentially available at this frequency band. Achieving these gains in practice, however, requires dealing with the severe propagation characteristics of high-frequency signals. To combat the enhanced path-loss in outdoor mmWave links and to provide sufficient received signal power, directional beamforming with large antenna arrays needs to be deployed at both the base station and mobile users. Further, when the objective is to send multiple data streams (instead of one stream), then more sophisticated precoding (instead of beamforming) schemes for large antenna systems need to be designed and employed. Unfortunately, the high cost and power consumption of the mixed-signal components makes dedicating a radio-frequency chain per antenna and performing the precoding processing in the baseband, the default solution in microwave systems, infeasible. Moreover, the design of the precoding matrices is usually based on complete channel state information, which is difficult to achieve in mmWave due to the large number of antennas and the small signal-to-noise ratio before beamforming. These challenges urge the need to develop new precoding and channel estimation algorithms tailored for mmWave systems.

Exploiting the sparse-nature of mmWave channels and leveraging tools from compressed sensing, WNCG graduate student Ahmed Alkhateeb, WNCG alumnus Omar El Ayach, Delft University Professor Geert Leus, and WNCG Professor Robert W. Heath Jr. developed channel estimation and hybrid analog/digital precoding algorithms for mmWave systems. The proposed algorithms estimate multi-path mmWave channels adaptively using different beamwidth beam patterns. Employing a hybrid analog/digital transceiver architecture, a novel scheme for constructing multi-resolution beam patterns was developed. The proposed channel estimation algorithms were shown to efficiently estimate the mmWave channels while requiring small training overhead, and while considering practical hardware constraints on the analog precoding devices. The channel estimation algorithms were also analytically evaluated, and lower bounds on its performance were derived.

This research funded by the National Science Foundation under Grant numbers 1218338 and 1319556 and a gift from Huawei.

Paper 1: Channel Estimation and Hybrid Precoding for Millimeter Wave Cellular Systems