Jukka Talvitie*, Toni Levanen*, Mike Koivisto*, Tero Ihalainen†, Kari Pajukoski†, and Mikko Valkama*
* Tampere University of Technology, Finland
† Nokia Bell Labs, Finland
Providing high-capacity radio connectivity for high-speed trains (HSTs) is one of the most important use cases of emerging 5G New Radio (NR) networks. In this article, we show that 5G NR technology can also facilitate high-accuracy continu- ous localization and tracking of high-speed trains. Furthermore, we describe and demonstrate how the NR network can utilize the continuous location information for efficient beam-management and beamforming, as well as for downlink Doppler precompensation in the single-frequency network context. Additionally, with particular focus on millimeter wave (mmWave) networks, novel concepts for low-latency intercarrier interference (ICI) estimation and compensation, due to residual Doppler and oscillator phase noise, are described and demonstrated. The provided numerical results at 30 GHz operating band show that sub-meter positioning and sub-degree beam-direction accuracies can be obtained with very high probabilities in the order of 95-99%. The results also show that the described Doppler precompensation and ICI estimation and cancellation methods substantially improve the throughput of the single-frequency HST network.
Published in IEEE Communications Magazine in September 2019
Videos and Demonstrations
Positioning of a high-speed train
In this demonstration, we illustrate the positioning performance of a high-speed train (HST) in a realistic 5G new-radio (NR) network building on the 3GPP specifications and 5G NR numerology. The velocity profile of the train is shown in the south-east corner, whereas the whole train trajectory is shown in the north-east corner together with the corresponding zoomed and more detailed illustrations on the left. Due to the geometry of the considerd HST scenario, the 95% confidence ellipse is relatively large in the perpendicular direction of the track when the remote radio heads (RRHs) are far away from the train, and hence, majority of the positioning error is stemming from this direction. However, the overall positioning performance is extremely high throughout the trajectory.