Overview
Through the use of Intermodal Transport Control Systems (lTCS) and passenger information systems, there are more and more digital systems in public transport, which have a high information potential. For scheduling tasks as well as for information services, it is not only the knowledge and transmission of the current traffic situation that is relevant, but also a reliable forecast of arrivals. This allows for improved real-time travel planning and assistance, including travel arrangements, and passenger information. Despite all the qualities achieved, the forecast quality is still unsatisfactory, especially in networked systems with a dense sequence of driving sequences and strong IV influences. This applies both to the forecasts for the time of arrival and the reliable prediction of the order in which vehicles arrive at the stop, which is important for passengers with limited mobility.
Until now, actual data for a forecast period has been extrapolated unchanged to the subsequent stop points without having to carry out forecasts for a delay recovery or increase in delay
a reliable time of arrival (order at stops) and - disposition measures for connection. With increasing forecasting horizons, the probability of occurrence and reliability have thus so far decreased significantly. Two main influencing factors are first that current timetable situations are often only updated without taking into account the reserves not included in the timetable; this tends to lead to an excessively long journey time.
Moreover, external influences caused by road traffic or unavailable timetable routes in the rail network tend to lead to an underestimated travel time estimate.
This project aims at investigating the possibilities and methods for increasing real-time quality in a pilot application in a metropolitan region. Therefore, an ITCS-based forecasting methods (e. g. historical-based method vs. timetable-based) will be investigated, further developed and tested in a pilot test for service quality and customer acceptance.
First of all, a survey of models and procedures used to date for predicting intermodal travel chains in public transport, with special attention to the effects on connection security and alternative routes is constructed. Then, the essential forecast and application cases as well as classification of the forecast quality
are defined.
Moreover, an analysis of alternative forecasting methods for the different forecast cases will be constructed.
The project also focuses on the investigation and assessment of various mathematical forecasting methods, taking into account the requirements analysis.
Furthermore, practical applicability and "automatic" scalability (through daily operation data) will be considered.
Then, an analysis of the effects of MRP functions on the forecast input values will be conducted and a selection procedure and evaluation of forecasting functions for practicable applicability will be done.
Finally, a test field, a trial and evaluation process is defined.
Finally, a report with recommendations for action will be provided.