A Fast Dictionary Learning Algorithm for CSI Feedback in Massive MIMO FDD Systems

In a massive multiple-input multiple-output (MIMO) frequency division duplex (FDD) system, it is required to compress the channel state information (CSI) and feed it back to base station (BS). In this paper, we primarily focus on the compressive sensing (CS)-based feedback design and propose a fast dictionary learning (FDL) algorithm to update the singular vectors of matrices in the K-singular value decomposition (K-SVD) algorithm. The proposed FDL algorithm is a variation of the existing K-SVD algorithm with low computational complexity. Simulation results also show that the proposed method’s estimated channel has better normalized mean-squared error (NMSE) performance than the estimated channel using a traditional Discrete Fourier transform (DFT) dictionary. Also, the proposed method’s estimated channel has comparable performance with the K-SVD algorithm but with reduced computational complexity ranging from 18% to 45%, which is significant.