Millimeter Wave Channel Estimation for Lens based Hybrid MIMO with Low Resolution ADCs

TL;DR

This paper introduces an EM-based mmWave channel estimator for lens-based hybrid MIMO systems with low-resolution ADCs, improving performance while reducing power and hardware complexity.

Contribution

It proposes a novel EM-based channel estimation method tailored for lens-based hybrid MIMO with low-resolution sampling, enhancing accuracy over traditional approaches.

Findings

Outperforms conventional EM in medium to high SNR

Provides increased antenna gain with lens-based beamforming

Reduces power consumption with low-resolution ADCs

Abstract

The high path loss associated with millimeter wave (mmWave) frequency communication can be compensated by large scale antenna arrays such as multiple-input multiple-output (MIMO) systems. The hybrid beamforming architecture which uses fewer radio frequency chains is implemented to reduce power consumption and hardware complexity, while still supporting multi-stream communication. We propose an efficient expectation-maximization (EM)-based mmWave channel estimator for a lens-based hybrid MIMO system with low resolution sampling at the receiver. The lens-based beamformer is investigated to provide increased antenna gain and reduced implementation complexity as the conventional beam selection network is excluded. Low resolution sampling at the analog-to-digital converters is implemented for reduced power consumption. The proposed solution with a robust maximum a posteriori estimator based on the EM algorithm performs better than the conventional EM approach and minimum mean square error baselines in medium to high signal-to-noise ratio regions.