8.0 Summary of Arterial Data Gaps
In this analysis, several information gaps have been identified:
- Traffic signal systems represent the majority of real-time data collection in most arterial transportation systems.
- Most traffic signal systems do not support transferring data to a central system for analysis or sharing with other systems.
- Most traffic signal systems do not collect the types of data required for ICMS.
- Most traffic signal systems do not support the data collection granularity required for ICMS. Speed, volume, and occupancy data are needed per lane on a 1-2 minute collection cycle. Many signal systems aggregate sensors across multiple lanes (as was illustrated in Figure 5) and most aggregate data into hourly summaries before uploading the data to central servers.
- Modification of traffic signal systems to provide the desired data usually requires modifications to the sensor connections, to the controller, and to the controller software. This may not be feasible in many corridors.
- Alternative data collection methods are limited to providing speed or travel time data which cannot easily be used to derive critical capacity factors such as traffic volume, occupancy, headway, and density.
- Speed and travel times are well suited to public dissemination, but speed and travel times do not predict capacity problems; they only identify the location once the capacity failure has occurred.
- The granularity of AVL and AVI data may not be a good match for operational needs. The travel time and speed data from these systems can represent 5-15 minute or longer time spans. Both techniques also offer limited coverage of the arterial network.
- The cell phone probe data has not had a good record of providing detailed traffic data on small arterial segments. This does not mean that the data is not useful, but that it may be useful in different ways from more traditional traffic sensor systems. The statistical methods being explored with the DUAP project may provide additional useful ways to use the cell phone and other probe data.
From the AMS portion of the Initiative, it will be critical to learn which types of data are essential to modeling the corridor, and what granularity of data is required for the AMS tools to be accurate and useful. From the pioneer site deployments, it will be important to learn which types of data gathering measures are effective and produce useful information for operations as well as the public.
The ICMS projects each have a unique modeling approach and different goals for how the modeling part of the ICMS will integrate with the operational and planning aspects of their corridor management strategy. This diversity should provide good insights into what can and cannot be done with arterial data to manage corridor capacity.
Researchers are evaluating performance measurements for traffic signal systems through a variety of approaches. While there is not a clear agreement on what performance measures should be used, there is a growing agreement that improvements in the analysis, modeling, and decision support capabilities for arterial systems will require some basic improvement to the data that is currently available from a controller:
- Data collected or calculated in the controller should be time-stamped and sent to the central system, not averaged or discarded.
- Firmware should be modified to collect phase, overlap, control mode, detector, and other event information at a significantly finer time resolution. (Controllers typically scan input data at 0.1s intervals and some researchers are suggesting sending 1.0s data updates to the central server.)
- Typical detector configurations may need to be modified to allow finer resolution of intersection volumes by approach and exit. Occupancy and volume need to be collected on a lane-by-lane basis.
- Additional sensors or modifications to existing sensor configurations may be needed to provide improved speed data (either in the form of travel time or actual measured vehicle speeds)
If it is impractical to implement these changes in existing controllers due to processor, memory, or firmware limitations, it may be feasible to implement separate data acquisition modules that share access to the primary detection devices.
It may be necessary to investigate new ways of collecting data from intersections using the existing sensors, but bypassing the signal controllers. It may be possible to piggyback COTS data acquisition equipment on the existing sensors without disruption to the signal system. This would allow intersections with older controllers to participate in the ICMS data collection at a lower cost than replacement of the signal system. Adding processor power to intersections may also be less risky than modifying controllers to collect more data than the original controller specifications required.
