A simulation study indicated that integrating traveler information with traffic and incident management systems in Seattle, Washington could diminish delay by 1 to 7 percent, reduce stops by about 5 percent, lower travel time variability by 2.5 percent, and improve trip time reliability by 1.2 percent.
Date Posted
06/15/2001
Identifier
2007-B00359
TwitterLinkedInFacebook

ITS Impacts Assessment for Seattle MMDI Evaluation: Modeling Methodology and Results

Summary Information

This study examined the impacts of integrated advanced traveler information services (ATIS), advanced traffic management systems (ATMS), and incident management systems (IMS) on a mixed freeway/arterial corridor in north downtown Seattle. The regional and corridor level impacts of ITS were captured using the Mitretek Systems Process for Regional Understanding and EValuation of Integrated ITS Networks (PRUEVIIN). EMME/2 was used as the transportation planning model, and INTEGRATION-1.5 was used to generate simulations. The performance measures analyzed included the following: near-term peak period delay reduction, travel time reliability, changes in regional mode choice, corridor travel throughput, fuel consumption, and emission rates.

Measuring ITS impacts over a range of conditions was the key element to an accurate representation of annualized impacts. Accordingly, the simulation was exercised through a series of 30 scenarios. Each scenario represented a particular combination of weather impacts, travel demand variation, as well as variations in incident patterns and accidents. The scenarios were calculated from cluster analyses of traffic flow data (for variations in day-to-day travel demand) and weather/incident impacts (taken from historical archives). Each scenario had a probability of occurrence. The scenarios taken together comprised a representative year of operation.

The evaluation of ITS was broken down into a collection of four groupings: ATMS, ATIS, IMS, and Integration.

  • ATMS included projects that archived and consolidated arterial traffic data from a number of sources and compiled the information at a central location.
  • ATIS was comprised of a collection of pre-trip and en route information services that provided current congestion information based on real-time Washington State Department of Transportation (WSDOT) freeway detector data.
  • IMS attempted to improve detection, response time, and freeway system efficiency under incident conditions.
  • Integration contained only one project (ITS Backbone), which collect data from arterial sensors and integrated ATIS with traffic signal control. Both regional and corridor level simulations were used to evaluate the potential "big-picture" impact of integration on regional travel.

Each experiment was compared to a uniform baseline case representing long-standing traveler information services and traffic management systems in the Seattle corridor. For example, commercial traffic reporting and ramp metering control on I-5 were considered as elements of the baseline case.

Network efficiency impacts included data collected for all vehicles with trips starting in the north corridor between 6:15 AM and 9:00 AM.

Average delay was calculated as the difference between the average travel time in each scenario and free-flow travel times (50 percent of average demand, no accidents in the system, good weather).

Delay reduction was calculated by expressing the difference in average delay between the baseline case and the experimental case as a percentage of baseline average delay.

Throughput was determined by measuring the number trips started between 6:15-9:00 AM and ended before the end of the 9:30 AM peak period. Delay reduction and throughput measures were calculated for each scenario. An annualized figure was then calculated by computing a weighted average of across all scenarios.

System coefficient of variation was calculated by examining the variability of travel time for similar trips in the system taken across all scenarios. This statistic was considered an indicator of travel time reliability in the corridor.

Energy estimates were calculated as total liters of fuel consumed, total hydrocarbons (HC), Carbon monoxide (CO), and Nitrogen oxides (NOx) emissions. Virginia Tech used Link-level speed and stop data derived from simulations to calculate energy consumption and emissions during peak network activity.

FINDINGS (Mobility)

ATIS Experiment

The ATIS experiment attempted to capture the near-term impacts of increased utilization of various traveler information services such as Microsoft (TM) Sidewalk Traveler Information, Etak/Metro Traffic Control Traveler Information, Fastline Hand-held PC, Traffic Channel on Cable TV, WSDOT Web Page and Traffic Telephone Information Line. The model discriminated between pre-trip or en route information, the coverage area, and the level of message detail. For experimental purposes, the ATIS pre-trip information user rate was estimated at 6 percent.

The traveler information services available in the baseline case provided incident, construction, and other emergency road closure information on radio, TV, variable message signs and highway advisory radio as they have been provided in Seattle for many years. In this baseline case, route choice decisions were made under greater uncertainty about the delays associated with incidents, weather, or recurrent bottlenecks. However, in the ATIS case, visual displays of I-5 freeway travel congestion were provided to enable travelers to more effectively gauge estimated travel times.

  • The inclusion of ATIS with baseline traveler information slightly improved overall system performance . The AM peak period total system delay in the north corridor was reduced by 1.5 percent. In addition, travel was more reliable as travel time variation was reduced by 2.5 percent. The impact of ATIS was highest in scenarios with poor weather, heavy demand, freeway accidents, or a combination of these.
  • Crashes were reduced by 0.6 percent. The expected number of fatal crashes over a ten year period was reduced by 0.4 percent .

ATMS Experiment

The ATMS experiment attempted to capture impacts of re-timed signals along two major arterials in the north corridor of Seattle. This prospective re-timing was based on detector data collected along arterial routes and archived at the northwest transit management center. The baseline signal timing plan was limited to fixed-timing plans that were optimized for peak period flow.

  • Total AM peak period delay in the north corridor was reduced by 7.0 percent (annualized reduction) with and average of 0.2 percent more vehicles entering and leaving the network during peak periods. Travel was slightly more reliable as travel time variation decreased by 2.1 percent.
  • Overall, the expected number of crashes decreased by 2.5 percent, and the total number of fatal crashes projected over a ten-year period decreased by 1.1 percent. This reduction was attributed to a shift from lower-speed travel to higher-speed travel. (Travel speed increased from 32 to 40 kilometers per hour (kph) to 60-80 kph).

IMS Experiment

In this experiment, system level impacts reflected the ability of the highway patrol, WSDOT, and emergency medical service providers to coordinate incident response. Relevant MMDI projects associated with this included Regional Video Sharing, Incident Information Capture, On-Scene Incident Video, and Emergency Operations Center Coordination. Reaction to an incident was characterized by detection time, response time (time to getting the first unit to the incident site), and incident clearance time.

Major disruptions on the freeway when combined with heavy demand or snow generated the most significant impact. However, the overall system impacts were limited since benefits were highly concentrated among users traveling to particular facilities at particular times.

Arterial Data for ATIS Integration Experiment

This experiment modeled the integration of data from arterial loop detectors (along routes SR99 and SR522) with data provided by the WSDOT freeway-based ATIS internet web site and other media. The baseline case assumptions were the same as with the ATIS and ATMS experiments. No changes to existing traffic signal controls along the two arterials were modeled. The only change was that users of ATIS were given the ability to consider real-time estimates of congestion on the two arterial routes (SR99 and SR522) as well as Interstate I-5 when making travel decisions. It was assumed that arterial data was updated every 15 minutes and provided a combined estimate of both link travel time and intersection delay.

  • The provision of arterial data roughly tripled the overall system impacts of ATIS in the north corridor. Vehicle hours of delay were reduced by 3.4 percent. Vehicle throughput was also higher, and trip time reliability was improved by 1.2 percent. Total travel time slightly increased, while stops decreased by 5.5 percent. Overall, it was clear that the provision of travel time estimates on the primary alternatives to I-5 in the north corridor allowed travelers to make more efficient route choice decisions. Patterns of use also changed; the total freeway to arterial diversion decreased when the arterial data appeared in the ATIS. The arterial ATIS reduced unwarranted diversions away from the main freeway when travelers were given more current and accurate estimates of arterial performance.
  • Overall, the expected number of crashes decreased by 1.0 percent with the overall number of fatal crashes projected over a ten-year period reduced by 0.3 percent.

Enhanced ITS Alternative Analysis

The Enhanced ITS Alternative was a prospective integrated deployment of ATIS and ATMS. The experiment featured an improved signal coordination system along routes SR99 and SR522, and a 6 percent user base for estimating Interstate I-5 freeway congestion, and SR99 and SR522 travel times. It was noted that this experiment was different that the other experiments in this study since both the regional and sub-area models (PRUEVIIN) were analyzed to estimate travel demand changes in response to system capacity improvements.

  • Regional level impacts were relatively small, however, sub-area delay was reduced by 6.1 percent and the total number of stops decreased by 4.7 percent.
  • Overall, the expected number of crashes decreased by 1.9 percent, with the overall number of fatal crashes projected over a ten-year period reduced by 0.8 percent. This increased crash rate was attributed to increased travel speeds and the additional trips on the network.