With the increasing demand for passenger and freight transport and the need to reduce greenhouse gas emissions, many countries are turning to railways. To enhance efficiency and safety, innovative operational concepts like virtual coupling and autonomous trains are being developed, which rely on reliable, low-latency communication. Future communication standards aim to improve railway communications and facilitate dependable train-to-train (T2T) communication. Achieving the necessary reliability requires consideration of propagation conditions in radio system design, particularly for safety-critical applications where correlated error events are critical. Until now, suitable channel models for T2T communications were lacking. This dissertation introduces a wideband T2T channel model based on C-band measurements, following the first wideband T2T channel sounding measurement campaign. The dynamic environment and moving transmitter and receiver result in non-wide-sense stationary uncorrelated scattering. The analysis of time-frequency-variant data leads to the evaluation of quasi-stationarity regions and the derivation of time-variant spreading functions. Stochastic channel parameters are estimated for typical railway settings, with multipath components isolated and tracked. A new metric, scattering loss, is introduced to assess the impact of elements on multipath fading. A geometry-based stochastic channel model (GSCM) for T2T c
