Using co-simulation and Time Signal at Red (TSAR) to Determine Impact of Driver Behavior on Rail Network Performance

Railway performance is hindered by reactionary delays due to interference between trains. On congested networks, driver performance is a key element affecting delays. Time Signal at Red (TSAR) is a performance metric with a fine level of granularity that provides insight into performance. We present an initial co-simulation of a section of railway in Great Britain, the Surbiton to Weybridge line near London, that includes high-fidelity physics models of trains, models of drivers with individual behavior characteristics, and a signalling model. The co-simulation generates TSAR data, which demonstrates the impact of driver behavior and interference between trains on performance. The simulated TSAR data is compared with real data, which shows qualitatively the potential of the approach. Future work will include improvements to the driver model, a quantitative validation of the simulation results with the real data, and extension of the co-simulation to include other models with complementary characteristics.