SVM-Based Optimization on the Number of Data Streams for Massive MIMO Systems
Shiguo Wang,Mingyue He,Rukhsana Ruby,Yongjian Zhang
DOI: https://doi.org/10.1109/jsyst.2021.3139991
IF: 4.802
2022-01-01
IEEE Systems Journal
Abstract:For millimeter-wave (mmWave) massive multiple-input and multiple-output (MIMO) systems, hardware cost and power consumption can be reduced via hybrid precoding, while keeping almost the same spectral efficiency (SE) as the fully digital precoding scheme. However, for different hardware configuration and channel environments, system SE varies with the number of transmitted data streams at the transmitter, and hence this should be determined primarily before hybrid precoding design. In this article, for single-user massive MIMO systems, a regression model based on the support vector machine technique is proposed, through which the optimal number of data streams can be obtained automatically based on the propagation characteristics of wireless channels, the number of antennas, and RF chains. To validate the correctness of the proposed model, simulation results are presented under various parameter settings.
computer science, information systems,telecommunications,engineering, electrical & electronic,operations research & management science