DDL-Based Sparse Channel Representation and Estimation for Downlink FDD Massive MIMO Systems.

Jianqiao Chen,Xi Zhang,Ping Zhang
DOI: https://doi.org/10.1109/icc40277.2020.9148996
2020-01-01
Abstract:We address the problem of sparse channel representation for downlink channel estimation in multi-user frequency division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems. Existing methods typically adopt discrete Fourier transform (DFT) matrix as a sparse basis to represent sparse channels. However, a sparse basis constructed through dictionary learning method has proven to have strong sparse channel representation. In this work, we develop a discriminative dictionary learning-based sparse channel representation for downlink channel estimation in a multi-user FDD massive MIMO systems. Considering partially shared support between near users, we present a new discriminative dictionary learning (DDL) method for sparse channel representation, based on which the channel estimation scheme is developed. Compared with learning a shared dictionary for all users, it can provide a better representation, thus improving the performance of recovery in the compressive sensing process. Numerical results demonstrate the superior performance of discriminative dictionary as compared to the shared dictionary in terms of normalized mean square error (NMSE) and symbol error rate (SER).
What problem does this paper attempt to address?