Intelligent Engineering Systems through Artificial Neural Networks
18 Predicting Travel Time along Arterials with More Sustainable Methods
Download citation file:
- Ris (Zotero)
- Reference Manager
Predicting travel time along arterials is a complex dynamic problem. Most complexity derives from traffic interruptions at signalized intersections, viewed as decision points with significant uncertainty. Accurate but costly and unsustainable prediction can be obtained using sensors and probe vehicles. This research presents a more sustainable approach that draws on advanced computational paradigms and easily available data, to predict travel time along urban arterials. The notion of state-space of roadway traffic and state-space neural networks (SSNN) are used to predict travel times both under time-invariant turn (TIVT) and time-variant turn (TVT) shares using basic traffic data easily collectable with current field hardware.