Summary
International Conference on Emerging Technologies for Communications
2020
Session Number:SB1
Session:
Number:SB1-5
Prediction of Wireless MmWave Massive MIMO Channel Characteristics Based on Graph Attention Networks
Long Yu, Cheng-Xiang Wang,
pp.-
Publication Date:2020/12/2
Online ISSN:2188-5079
PDF download
Summary:
This paper proposes a procedure of predicting millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) indoor channel characteristics based on graph attention networks (GAT). We use the K-nearest neighbor (KNN) algorithm to construct the real channel measurement data into a graph dataset. Different from existing machine learning (ML) based channel characteristics prediction algorithms using all data points at the same time, we only use some data with high correlation to train our model in order to reduce complexity and number of iterations. Scenario parameters including transmitter (Tx) and receiver (Rx) coordinates, Tx-Rx distance, and carrier frequency are used to characterize the correlation between data points, while the output parameters are channel statistical properties, including the received power, root mean square (RMS) delay spread (DS), and RMS angle spreads (ASs). The predicted channel statistical characteristics can fit well with those of real channel, which indicates the effectiveness of proposed method.