Travel Time Messaging on Dynamic Message Signs - Houston, TX
Texas Department of Transportation
Table of Contents
2.0 Deployment Information
3.0 System Planning and Development
4.0 Data Collection and Processing
5.0 Travel Time Messaging
6.0 Public Outreach and Impacts
7.0 Issues Faced and How They Were Resolved
The Houston TranStar partnership was created in 1993 to coordinate and enhance surface transportation operations in Harris County, Texas. It is comprised of the four principal transportation and emergency management agencies in Harris County:
- Texas Department of Transportation (TxDOT)
- Metropolitan Transit Authority of Harris County (METRO)
- Harris County
- City of Houston
The Houston TranStar partnership is responsible for coordinating the planning, design and operation of transportation systems and emergency management functions, as well as the development and deployment of Intelligent Transportation Systems (ITS).
TranStar plays a vital role in the management of Houston-area traveler information systems. Traveler information is disseminated by four means: dynamic message signs (DMS), highway advisory radio (HAR), the Internet, and the local media. TranStar began posting travel times onto DMS in 1996 and currently posts travel times onto 81 DMS across more than 250 centerline miles of Houston area freeways. These travel times are based on data collected from the close to two million "EZ-Tag" Automatic Vehicle Identification (AVI) toll transponders currently circulating around the Houston metropolitan area. Data from these vehicle probes is collected at 232 supplemental reader stations and transmitted to the TranStar TMC for analysis. Reader stations are, on average, two-to-three miles apart, but not more than five miles apart.
The initial push behind the posting of travel time information onto DMS was the result of driver complaints concerning the posting of safety and other static messages to the DMS. As a result of these complaints, TranStar staff conducted an Internet-based survey to determine the types of information people wanted to see posted onto the DMS. Results of this survey indicated that drivers were primarily interested in seeing incident information and travel times. More recent surveys indicate that drivers now prefer that travel time information be posted above all other data.
In addition to posting probe-based travel times onto area DMS, vehicle speeds derived from the travel time system are used to support a traffic map for the greater Houston region (see Figure 1). This traffic map displays real-time traffic conditions and is used by TranStar staff for traffic management and incident detection purposes. It is available to the public via Houston TranStar's website: www.houstontranstar.org. Popularity of the website has resulted in heavy usage, especially during periods of inclement weather (e.g., tropical storms and flooding). As a result, TranStar has been forced to procure bigger, faster servers, as well as 10 T1 lines, with burstable capacity up to 100 T1 lines.
Communications between supplemental AVI tag readers and the TranStar TMC are mainly supported via dial-up modem over traditional phone lines. However, TranStar is currently in the process of converting to CDMA and GPRS wireless models. The reasoning behind this is that it is believed that use of these modems will improve O&M and cut ongoing communications costs in half. Almost all DMS are connected to the TranStar TMC via the existing fiber network.
TranStar's travel time system was initially manual in nature. One operator posted travel times to about 40 DMS every 15-20 minutes during peak travel times. TranStar eventually decided that continuation of a manually-driven travel time calculation/posting to DMS process would not be feasible over the long-term. In response, TranStar worked with the Texas Transportation Institute (TTI) to develop an automated travel time processor and the Southwest Research Institute (SwRI) to develop an automated DMS interface for posting messages to the signs. This fully automated system currently posts travel times to 81 signs every 10 minutes (5:30 am-7:30 pm), with some DMS updated more frequently depending on need, for an average of 4900 messages posted per day. If this level of service were to be provided via the manual calculation of travel times/posting of DMS messages, at least four operators would be needed per shift just to do this work.
One ongoing problem concerns the high rate of turnover in TMC operations staff. TranStar's managers recognize the urgency of this issue but haven't yet found the best way of dealing with it. Attempts have been made to document job requirements in order to help decrease the learning curve for new hires, but this can be difficult to do. Although all new operators get three months of training on all TMC equipment, it usually takes between one and two years before they are ready to make decisions themselves.
Although most initial capital costs were paid for by FHWA/TxDOT funds, recurring costs remain high. As a result, TranStar has sought financial support from other stakeholders to provide funds to support recurring costs related to system operations and maintenance. Operations and maintenance includes items such as paying for licensing of proprietary software and communications costs between the sensors, DMS, and TMC.
TranStar's long experience with posting travel time information onto DMS has led them to recommend that public agencies interested in implementing a system should "take a systematic, well thought out approach to getting started - corridor by corridor implementation and evaluate along the way." In other words, agencies should take baby steps to get to the end so that problems can be corrected before they become too large.
Other planning-related recommendations include:
- The public agency involved should do whatever it can to gain ownership of software/other proprietary material. Failure to do so could result in problems for the agency when system modifications are needed.
- Legacy systems will be difficult to automate as they likely work as part of different subsystems and applications. In order to avoid this problem, the agency involved should seek to implement standards-based procurements for the purpose of facilitating system interoperability.
TranStar's current travel time data collection system uses vehicles equipped with transponder tags as vehicle probes. The main source of vehicle probes is commuters using the "EZ-Tag" automatic toll collection system installed by the Harris County Toll Road Authority (HCTRA). Transponder tag readers are placed at 1 to 5 mile intervals along freeways and HOV lanes. Each reader senses probe vehicles as they pass a reader station and transmits the time and location of the probes to a central computer for analysis. As the probe vehicles pass through successive AVI readers, software calculates average travel times and speeds for a roadway segment. The averages are made available to software which provides the data for the Houston TranStar web site and for posting to the DMS. It should be noted that the TranStar travel time processor strips all individually identifiable information from recorded tag reads in such a way that they can track, but not identify, a vehicle as it travels across the data collection network. Figure 2 shows one of the AVI readers in Houston.
At first, TranStar operators used the CCTV network to regularly assess data accuracy by comparing travel times calculated by the TranStar system against the time it took individual vehicles to drive a given segment of roadway. However, once the system's stability was established, it was determined that such verification efforts would not need to be conducted as often. Even so, TTI staff can alert TranStar operators if a problem is detected with the travel time package, a sensor station, or data being collected for analysis.
Survey results indicated that drivers prefer a single phase message (steady, single page message). As a result, this is the preferred format for TranStar's DMS messaging. Double phase messages continue to be utilized, but only when highly important information needs to be conveyed. As the potential exists for data latency issues to crop up in a probe-based travel time data collection system, it was decided that the system would post both the travel time and the time at which that travel time was calculated (see Figure 3, below). As a result, TranStar staff consider their system's accuracy level to be close to 100 percent.
TxDOT's system is capable of posting travel time information to the DMS 24 hours per day, seven days per week. However, travel time messages are typically posted between 5:30 am and 7:30 pm daily, with updates every 10 minutes (more often at some locations). Travel times are also posted on DMS in any situation where conditions depart from free flow, even if it occurs outside the system's normal operational period.
TxDOT conducts all DMS messaging activities according to the following messaging hierarchy:
- Amber alerts
- Travel times
- Special events
- Safety campaigns
No quantitative results are currently available to indicate whether the posting of travel times onto DMS has had a positive impact on system performance, but survey results (see below) indicate that the provision of this information has resulted in alternate routing by area drivers. In addition, TranStar staff indicate that people have come to rely on the system and complain if it is not available. TranStar has also gained the support of Houston-area politicians, who make regular requests for the deployment of additional DMS that will expand system coverage for voters within their districts.
TxDOT has conducted a number of surveys to assess customer satisfaction and evolving driver needs, the most recent taking place between April-May 2004. Results of this survey included:
- 85 percent of the respondents changed their route due to information on a sign (of this 85 percent, 66 percent said the route change resulted in travel time savings, 29 percent were not sure).
- Great support was expressed for the DMS program - many respondents indicated that they depend on the information and react accordingly by changing routes.
- Types of alerts with perceived positive benefits were as follows
- Incident alerts - 93 percent
- Freeway travel times - 82 percent
- Real-time road work advisories - 81 percent
- Future road work advisories - 70 percent
- Severe weather information - 67 percent
- Amber alerts - 61 percent
- Special event information - 31 percent
- Time sensitive messages should not be older than 15 minutes.
- Many respondents stated that incident information is nice, but that they also need travel time information to better determine how the incident impacts their travel.
- Respondents also requested more signs, expanded coverage, and messages specific to their individual desires.
|Manual operation of the system||If the system were operated manually today, four operators would be needed per shift to maintain it. With system automation, little ongoing work is needed from TxDOT operators to run it - primarily spot checking of data/other maintenance in response to alerts from software package.|
|Complaints about safety messages on DMS||Used survey to assess what drivers really wanted and began posting travel time information.|
|Maintaining drivers' privacy||Strip individually identifiable data out of tag reads to make probe data anonymous in nature.|
|Ongoing O&M costs||TxDOT/FHWA funds paid for most initial capital costs, but recurring maintenance costs are high. Sought out other stakeholders within the project area to contribute to maintenance costs, ongoing software costs, and communications costs for talking with the DMS. TxDOT is in the process of switching from use of telephone lines to CDMA/GPRS modems - will improve O&M, as well as reduce ongoing costs.|
|Drivers don't understand DMS system||Posted messages on related website so that drivers could learn about the nature of the system. TxDOT continuously reviews comments posted to their website in order to keep track of trends in user opinion.|
|Turnover of TMC staff||Great deal of turnover in TMC operations staff. Documenting job requirements to help decrease the learning curve, but this is difficult to do. All new operators get 3 months of training on all equipment, but up to 1-2 years are needed before they are ready to make decisions themselves.|
|Integration with legacy systems||Need to procure equipment that works on same standards, so that they are more likely to work as part of different subsystems.|
- Public agencies should own software or it could lead to problems when modifications are required later on.
- Although more expensive initially, automated data collection, processing, and DMS systems provide significant long-term benefits (e.g., TxDOT immediately knows when any part of the system needs maintenance).
- Some of the biggest operations and maintenance issues surround communications costs between the TMC and DMS.
- Need to take a systematic, well thought out approach to getting started. TxDOT recommends a corridor-by-corridor implementation with evaluation along the way. It is best to take baby steps in order to make it possible to resolve problems before they get out of hand.
Federal Highway Administration
U.S. Department of Transportation
400 7th Street, S.W. (HOP)
Washington, DC 20590
Toll-Free "Help Line" 866-367-7487
Publication Number: FHWA-HOP-05-051