About the project
This project investigates the travel time reliability of trips involving public transport, so as to enable more accurate prediction of a traveller’s estimated time of arrival for inclusion in multi-modal journey planners.
Travel time in transport networks entails variability that can be an important source of time losses. This results in uncertainty when attempting to predict it, which is a concern for travellers, as it can cause early or, more critically, late arrival at the destination. Research to date has concentrated on estimating the reliability of car trips, with journey planners nowadays being able to provide a fairly accurate Estimated Time of Arrival (ETA) for any car trip, along with upper and lower bound estimates.
Nevertheless, this is not the case for trips involving public transport, whose reliability remains under-explored. This is because such trips are constituted of multiple segments, and for several of these, until recently, travel time data have either been proprietary and scarce (e.g. in-vehicle and dwell times), or did not exist at all (e.g. access/egress times, waiting times and transfer times). As a result, journey planners still largely rely on the limited information that is freely available (e.g. published schedules) and on crude and conservative estimates (e.g. assuming a certain walking pace).
The aim of this project is to perform a deeper exploration into the reliability of trips involving public transport, so as to enable more accurate prediction of a traveller’s ETA. Tasks will include:
- collecting and analysing data for the various components of public transport trips from relevant sources
- formulating travel time prediction models
- integrating travel time predictions in journey planning algorithms
- validating the models and algorithms through simulation and field experiments
