In the Washington DC metropolitan area, drivers who use route-specific travel time information instead of wide-area traffic advisories can improve on-time performance by 5 to 13 percent.
Date Posted
11/29/2005
Identifier
2005-B00285
TwitterLinkedInFacebook

A Comparison of Mobility Impacts on Urban Commuting Between Broadcast Advisories and Advanced Traveler Information Services

Summary Information

This study used an analytical technique called the Heuristic On-line Web-Linked Arrival Time Estimator (HOWLATE), to quantify the mobility benefits of listening to radio traffic advisories and compare them to those of a prospective traveler information service designed to provide door-to-door estimates of real-time travel times. The case study was conducted for the Washington, DC metropolitan area using route-specific travel times archived from the SmarTraveler.com web site, and radio traffic reports recorded from a local radio station.

The roadway network modeled in the study included 33 bi-directional roadway segments (18 freeways and 15 major arterials) and used an archive of link travel times reported by SmarTraveler.com to reconstruct changes in travel times on major roadways during peak periods. The travel time data input into the model were collected using an automated process to archive the real-time travel time estimates shown on the website every five minutes from 6:30 AM to 6:30 PM for each weekday between June 1, 2001 and January 17, 2002. The travel times collected were corrected for potential errors based on probe vehicle runs conducted on one arterial and one freeway. The broadcast radio data set input into the model included an archive of 4,410 broadcast advisories that were automatically recorded over 37 weekdays between June 1, 2001 and January 17, 2002. The radio reports were manually processed to assign levels of congestion to roadway segments based on the type and severity of advisories mentioned.

To compare the on-time arrival performance of seasoned commuters provided with personalized route-specific travel time information and seasoned commuters listening to radio advisory information, two sets of simulated paired (yoked) driving trails were conducted using the HOWLATE procedure. Each simulated pair had the same origin, destination, desired arrival time and habitual route, but the subjects were not restricted to leave at the same time. One subject in each pair made use of traveler information (i.e., radio or the personalized notification-based service) and the other did not.

The radio listener had the advantage of making route modifications en-route; however, the traffic information provided was vague and imprecise, and did not always cover his chosen route. The user of personalized route-specific travel time information was only provided with pre-trip information; however, the information provided was more detailed and relevant to his chosen route, enabling a more informed trip departure time choice and route selection.

RESULTS

The analysis showed that the simulated commuter receiving route-specific travel time information typically made more effective route and trip timing decisions than the simulated user of broadcast radio traffic reports. In fact, the commuter using radio reports recorded similar on-time reliability performance as the baseline control subject, who ignored all forms of traveler information. During the afternoon peak period, when travel time variability was higher, the radio listener recorded slightly lower travel disutility than the baseline control subject; however, during all periods of the day, the on-time reliability of the radio listener was worse than the baseline control subject.

Comparison of performance measures for the AM and PM Peak Periods:

Aggregate Trip Metrics
Non User
AM Peak
Travel-time Information User
AM Peak
Radio User
AM Peak
Non User
PM Peak
Travel-time Information User
PM Peak
Radio User
PM Peak
Trip Time (minutes)
34.6
34.2
34.6
34.9
34.8
34.9
Travel Disutility Cost ($)
2.59
2.52
2.8
2.73
2.52
2.63
On-Time Reliability
88%
99%
86%
95%
98%
93%
Early Trips
5%
16%
6%
24%
12%
12%
Late Schedule Delay (Minutes)
2.8
2.5
3.9
2.2
3.4
3.4
Early Schedule Delay (Minutes)
11.9
12.9
11.4
14.1
12.2
11.3
Goal Areas
Deployment Locations