Automated Analysis of Train Event Recorder Data to Improve Micro-Simulation Models

This paper describes how on board digital train event recorders have been developed in recent years. These devices allow one to collect very detailed data about train movements and signal status. The new Italian ATC SCMT on board subsystem is combined with the DIS (Driver Information System) that collects both kinetic behavior and all signal and balises messages. Unfortunately, this large amount of data is normally stored but not used except for failure and maintenance management. At the same time the use of micro simulation tools has been extended to large scale problems. A known problem exists in the calibration and validation of these models. In this paper a new tool is presented. This software allows one to analyze real life collected data, to perform very detailed analysis of train movements, pointing out speed depending on position and signal aspects, acceleration, braking curves and dwell time graphically and by means of parameters. Train behavior can also be connected to punctuality, to find out differences between on time and late running. This tool may be very useful for: large scale model validation, definition of the stochastic behavior of the system (travel time, dwell time, initial delay), calibration of braking and acceleration curves for various train types, acceleration percentage depending on different conditions. In other words, the software allows one to set up a link between real data and micro simulation models. The tool has been tested in the north-eastern part of Italy. In this case study, a significant precision increase in the stochastic simulation results has been reached.