Joint Channel Direction Information Quantization For Spatially Correlated 3D MIMO Channels

TL;DR

This paper introduces a novel joint quantization codebook for 3D MIMO channels that leverages Tucker decomposition to efficiently encode channel direction information, reducing redundancy and improving performance.

Contribution

It presents a Tucker decomposition-based codebook for joint CDI quantization in 3D MIMO channels, significantly reducing the required statistical information while maintaining performance.

Findings

Achieves the same quantization performance as rotated codebooks with less statistical info

Provides substantial performance gains over independent CDI quantization

Validates the approach through analytical and simulation results

Abstract

This paper proposes a codebook for jointly quantizing channel direction information (CDI) of spatially correlated three-dimensional (3D) multi-input-multi-output (MIMO) channels. To reduce the dimension for quantizing the CDI of large antenna arrays, we introduce a special structure to the codewords by using Tucker decomposition to exploit the unique features of 3D MIMO channels. Specifically, the codeword consists of four parts each with low dimension individually targeting at a different type of information: statistical CDIs in horizontal direction and in vertical direction, statistical power coupling, and instantaneous CDI. The proposed codebook avoids the redundancy led by existing independent CDI quantization. Analytical results provide a sufficient condition on 3D MIMO channels to show that the proposed codebook can achieve the same quantization performance as the well-known rotated codebook applied to the global channel CDI, but with significant reduction in the required statistical channel information. Simulation results validate our analysis and demonstrate that the proposed joint CDI quantization provides substantial performance gain over independent CDI quantization.