Empirical Study on Directional Millimeter-Wave Propagation in Railway Communications Between Train and Trackside

• Published in: IEEE journal on selected areas in communications 2020-12, Vol.38 (12), p.2931-2945
• Main Authors: Yu, Daizhong, Yue, Guangrong, Wei, Ning, Yang, Lin, Tan, Hongcheng, Liang, Dan, Gong, Youhua
• Format: Article
• Language: English
• Subjects: 5G mobile communication; Algorithms; Antenna measurements; Beamforming; Cellular communication; Cellular structure; channel modeling; Data models; Directional antennas; Empirical analysis; High speed rail; High-speed rail transportation; Loss measurement; measurement campaign; Millimeter waves; Millimeter-wave; Misalignment; path loss; Railway stations; Receivers; Signal strength; Solid modeling; Wave propagation
• ISSN: 0733-8716; 1558-0008
• DOI: 10.1109/JSAC.2020.3005488

The high mobility of future high-speed trains (up to 500 km/h) poses great challenges on the design of dynamic beamforming (BF) algorithms for millimeter-wave (mmWave) communications in high-speed rail (HSR) networks. Thanks to the linear structure of HSR cellular networks, the track of the train is almost predictable. Thus, the fixed BF, i.e., highly directional antennas with fixed pointing directions, can be employed as a low-cost solution to replace dynamic BF. However, the performance gap between dynamic BF and fixed BF should be evaluated to justify the use of fixed BF in HSR mmWave communications. In this paper, empirical studies are conducted based on the raw data obtained from extensive measurement campaigns. Two classic environments in railway traffic are considered: the traditional train station and the HSR tunnel. Analyses of the measurement results, including the received signal strength, power delay profile, root-mean-squared delay spread, and channel non-stationarity are presented. Then, based on the measured data, we generalize the widely-used close-in (CI) free-space path loss (PL) model so that the generalized model can characterize the PL in the HSR tunnel with a higher accuracy. Finally, the performance gap between perfect dynamic BF and fixed BF is evaluated based on the generalized model and the measured data. Our results show that the average throughput of dynamic BF is only 4% higher than that of fixed BF in the HSR tunnel, but 21% higher in the train station when severe beam misalignment is present.