4B.4 Elements for Predicting Road Conditions

Tuesday, 9 January 2018: 9:30 AM
Room 10AB (ACC) (Austin, Texas)
Brenda C. Boyce, Booz Allen Hamilton, Alexander, AR; and P. A. Pisano and G. Guevara

Transportation system management and operations (TSM&O) are on the cusp of dramatic changes due to increased availability of data and sophistication of models and systems supporting those operations. Intelligent Transportation Systems (ITS) are widely deployed and gather data about weather and traffic conditions from across the road networks. The imminent deployment of connected vehicles will bring an orders-of-magnitude increase in data availability. Traffic and road condition predictions need and are powered by these data, and the accuracy and reliability of the models improve with the increase. This convergence of opportunities presents potential for operational improvements in safety and mobility. The Federal Highway Administration’s (FHWA) Road Weather Management Program (RWMP) is conducting research into an integrated model for road condition prediction (IMRCP) to investigate and capture that potential.

Transportation agencies currently have a variety of ITS collecting information to support TSM&O, but the integration of the data into operation support systems is selective. Predictive capabilities are typically limited to generalized weather forecasts and traffic predictions based on historical patterns. Impacts of other factors—hydrology, incidents, and work zones, for example—are even less likely to be routinely integrated with operational forecasts and plans. In many cases, winter maintenance teams have access to maintenance decision support systems (MDSS) based on forecast conditions, but these are not generally shared with traffic operations. Traffic management center (TMC) operations typically have strategies for dealing with abnormal events as they occur, but generally do not explicitly consider forecast conditions. Unusual and exceptional events (e.g., small stream flooding or coastal storm surge) are responded to, but not necessarily anticipated. Traveler information presents a distillation of current information, but seldom provides end user travel decision support.

The IMRCP provides an integrated view of forecast road weather and traffic conditions for a given road network. The IMRCP model draws input from hydrological and traffic data sources and a diverse set of weather events to generate estimates of current conditions and forecasts of future conditions. Forecast outputs are available through a web interface on maps, in reports, and in subscriptions.

Traffic data sources such as advanced transportation management systems (ATMS) provide volumes and speeds, freeway control and traffic signal operations data, incident reports, and plans for work zones and special events. Current and forecast atmospheric and hydrological conditions are drawn from National Weather Service sources. State and local agencies provide specialized road weather conditions such as pavement temperatures. Data collected from the various sources are indexed, stored and archived in a heterogeneous data store.

While atmospheric, hydrological, work zone and special event forecasts are taken from external sources, the IMRCP synthesizes road weather and traffic condition predictions with embedded best-in-class forecast models. In the current implementation, road weather conditions are estimated across the network using field measurements of conditions, and predicted from atmospheric forecast conditions using the METRo model. Current traffic conditions are similarly estimated from detector stations and demand models, and predicted from road weather, incident and demand forecasts using the TrEPS/DYNASMART model.

The IMRCP provides prediction data on web-based maps, reports, and subscriptions. The map enables users to select map layers for roadway, regional atmospheric and point-specific alert data. A traffic-focused map, for example, could display traffic, precipitation intensity and traffic incident alerts. Available map layers also include select route travel times, weather radar, NWS advisories and warnings and local road condition alerts, among many others. All data are available in reports and subscriptions that can be accessed by other systems. Maps and reports can also display archived data.

A portion of the Kansas City metro area along a congested interstate corridor and surrounding arterials has been used for a demonstration study and evaluation area. The Kansas City area is subject to highly variable weather conditions and local recurring congestion typically of U.S. urban/suburban settings. The I-435 corridor along the southern part of the metro carries heavy commute traffic in both directions and for much of its length runs along a stream way with historically significant flood risk. The corridor is well-instrumented for traffic, weather and hydrology.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner