Study of BEM-Type Channel Estimation Techniques for 5G Multicarrier Systems

TL;DR

This paper explores BEM-type channel estimation techniques for 5G multicarrier systems, demonstrating a low-complexity method that performs close to LMMSE without requiring channel power delay profile knowledge.

Contribution

The paper develops a BEM-based channel estimator with a strategy for covariance matrix estimation, suitable for both OFDM and GFDM systems in 5G.

Findings

BEM-type estimator achieves near-LMMSE performance

Method is robust and low-complexity

Effective in both OFDM and GFDM systems

Abstract

In this paper, we investigate channel estimation techniques for 5G multicarrier systems. Due to the characteristics of the 5G application scenarios, channel estimation techniques have been tested in Orthogonal Frequency Division Multiplexing (OFDM) and Generalized Frequency Division Multiplexing (GFDM) systems. The orthogonality between subcarriers in OFDM systems permits inserting and extracting pilots without interference. However, due to pulse shaping, subcarriers in GFDM are no longer orthogonal and interfere with each other. Due to such interference, the channel estimation for GFDM is not trivial. A robust and low-complexity channel estimator can be obtained by combining a minimum mean-square error (MMSE) regularization and the basis expansion model (BEM) approach. In this work, we develop a BEM-type channel estimator along with a strategy to obtain the covariance matrix of the BEM coefficients. Simulations show that the BEM-type channel estimation shows performance close to that of the linear MMSE (LMMSE), even though there is no need to know the channel power delay profile, and its complexity is low.