Exploring the Potential of Neural Networks for Bicycle Travel Time Estimation
2020; Springer Nature; Linguagem: Inglês
10.1007/978-3-030-55973-1_60
ISSN1867-4941
AutoresGiulia Reggiani, Azita Dabiri, Winnie Daamen, Serge P. Hoogendoorn,
Tópico(s)Transportation Planning and Optimization
ResumoAReggiani, GiuliaDabiri, AzitaDaamen, WinnieHoogendoorn, Serge tool for travel time estimation of cyclists approaching a traffic light can monitor level of service of intersections in bike crowded cities. This work represents a first step in developing such a tool. Neural NetworkNeural networks models are evaluated on how they perform in estimating individual travel time of cyclists approaching a signalized intersectionSignalized intersections. Based on simulated scenarios, in cities with low bicycle levels (deterministic scenario), Neural Networks are good travel time estimators whereas, in places with high bike volumes (where cyclists depart with a discharge rate) information on queued cyclists is crucial for travel time information.
Referência(s)