Abstract
This research examines the short-term travel time prediction problem on a link level basis (section of freeway between adjacent traffic sensors). The prediction problem can be abstracted as one of finding a mathematical relationship between the response variable (actual vehicle travel time) on the explanatory variables (detector measurements of speed, flow and occupancy). We use a neuro fuzzy logic approach which gives better extrapolation capability to the models and performs significantly better than purely data driven models such as neural network. Further analysis showed that the inherent variability in travel times limits the accuracy that the models can achieve.
Original language | English (US) |
---|---|
State | Published - 2006 |
Externally published | Yes |
Event | 2006 IIE Annual Conference and Exposition - Orlando, FL, United States Duration: May 20 2006 → May 24 2006 |
Conference
Conference | 2006 IIE Annual Conference and Exposition |
---|---|
Country/Territory | United States |
City | Orlando, FL |
Period | 5/20/06 → 5/24/06 |
Keywords
- Dual loop detectors
- Membership function
- Neural network
- Neuro fuzzy logic
ASJC Scopus subject areas
- Industrial and Manufacturing Engineering