A large-scale mimo channel state information compression and reconstruction method based on deep learning attention mechanism

Abstract

The invention discloses a massive MIMO channel state information compression and reconstruction method based on a deep learning attention mechanism, comprising the following steps: Step 1: perform DFT transformation on an input channel matrix; Step 2: construct a DS-NLCsiNet model; Step 3: Train the model to obtain the parameters of each layer of the model; Step 4: Perform inverse DFT transformation on the output of the model. The present invention is a large-scale MIMO channel state information compression and reconstruction method based on a deep learning attention mechanism, so that the correlation of long-distance channel information can be efficiently extracted, and at the same time, deeper features can be extracted with fewer parameters, which greatly improves the channel. Efficiency of matrix feature extraction.

Description

technical field [0001] The present invention relates to the technical field of massive MIMO systems, in particular to a method for compressing and reconstructing massive MIMO channel state information based on a deep learning attention mechanism. Background technique [0002] Massive MIMO (multiple-input multiple-output) system has been recognized as one of the key technologies of 5G wireless system. This technology greatly increases wireless communication by configuring a large number of antennas at the base station to form multiple independent channels in the spatial domain. system throughput. However, in a frequency division duplexity (FDD) MIMO system, the channel does not have reciprocity, and the downlink CSI obtained at the user end needs to be sent to the base station through the feedback link. resource cost. In addition, as the number of antennas increases greatly, traditional feedback reduction schemes such as limited feedback algorithms based on quantization and...

AI Technical Summary

Problems solved by technology

[0005] However, traditional models generally only use 3x3 or 5x5 convolution kernels, feature extraction is limited to a certain local neighborhood, and their receptive field is increased through continuous iteration, but this iterative process is very inefficient and relies on long-range The extraction effect is poor, requiring huge computing resources, and the output of each layer is only passed to the next layer, requiring more parameters, and the efficiency of feature extraction is low

When there are many layers, there will be gradient dispersion, which greatly reduces the transmission efficiency and accuracy

Images