Characterization of Time-Variant Wireless Channels in Railway Communication Scenarios

We present results from a train-to-infrastructure (T2I) and train-to-train (T2T) measurement campaign where we perform single-input-single-output channel sounding with a 150 MHz bandwidth, allowing us to resolve multi-path components with high precision. We analyze the measurements done for both T2I and T2T communication links, with trains going at speeds of maximum 20 km/h approximate to 5.56 m/s. Using low velocities allows us to run two trains behind each other on the same rail track with very short distances (0 < d < 200 m). The analyzed data combined with GPS and video recordings serves for describing the geometry of the environment and developing a physical geometry-based stochastic channel model that enables virtual channel emulation for scenarios with trains moving at high velocities ( 0 < v < 300 km/h approximate to 83.33 m/s), but short distances. Characterizing the channel through empirically obtained data in such scenarios is still impossible in practice due to safety issues. Nevertheless, it presents a key step towards ultra-reliable low-latency link research for 5G use cases in the high-speed railway domain. We show that neighboring traffic and surrounding infrastructure has a big impact on the delay spread and that at extremely close distances, path loss variation depends on used antenna polarization.