Principal-Component-Analysis-Inspired Channel Feedback Framework: Sorting-and-Sampling and Interpolation-and-Rearrangement

Abstract

In this letter, we propose a compression method to feed back the highly correlated large-size channel-state information (CSI) of massive multiple-input multiple-output systems. The proposed compression method is based on principal component analysis (PCA), which can reduce high-dimensional data to a smaller dimension by exploiting the correlations in the data. Motivated by PCA, to further reduce the feedback overhead and to reduce the computational complexity, we newly designed a transformation matrix that sorts and samples CSI for feedback. Accordingly, a transmitter interpolates and rearranges the feedback signal to reconstruct the CSI. The proposed sorting-and-sampling and interpolation-and-rearrangement (SSIR) can be readily applied for high-dimension reduction in any domain, such as the spatial (antenna), frequency, and time domains. Numerical results verify the compression efficacy of the SSIR feedback method.