Infrastructure Consortium Setting "Stretch Goals" for Intersection Collision Avoidance
IVsource.net
19 September 2000

• “sensor-friendly” tags or other markings that can improve the performance of some in-vehicle sensors;
• point communication systems (like dedicated short range communication, or DSRC) that allow information to be passed to and accepted from vehicles;
• global positioning systems (GPS) that will allow extremely accurate location of vehicles; and
• roadway databases that, when combined with GPS locations, can provide detailed geometric and operational descriptions of the roadway that can be used to make various IVI services more effective.

Considerable infrastructure investments will be required to make these systems a reality, yet experts anticipate that benefits in fact will outweigh costs if high severity collisions -- such as those occurring at intersections -- can be avoided.

IC Has Geographically Diverse Makeup

Virginia, California, and Minnesota DOTs form the current membership of the IC, along with FHWA.   The IC began in 1998 as a federally designated pool fund, with these states  as members.  The federal government sees the Consortium as the focal point for innovative research in cooperative IVI services, with a mission to serve as stakeholders for all state and local governments with respect to cooperative systems, and to promote the deployment of cost-effective cooperative IVI services.

"Just as IVI has partnered with the vehicle manufacturing industry to jointly conduct research that can be shared on a “pre-competitive” basis between manufacturers, IVI will partner with the Infrastructure Consortium to ensure that the research program for cooperative systems will provide results that can be accepted and deployed by governments.  The Infrastructure Consortium will provide a unique opportunity to allow partnerships between the vehicle manufacturing industry and the infrastructure providers so that cooperative vehicle-highway systems can be researched, evaluated, manufactured, deployed, operated, and maintained," states the summary document.

Significant Progress Made at Minneapolis Workshop

With Slater's "stretch goal" ringing in their ears, the Infrastructure Consortium’s initial focus is to develop infrastructure-based intersection collision avoidance systems.  The August workshop, held in Minneapolis, focused on establishing general objectives, defining specific crash scenarios that could be addressed, identifying candidate intersection collision avoidance concepts, and specifying research tasks needed to develop such systems.

The Infrastructure Consortium members agreed that the deployment of infrastructure-based intersection collision avoidance systems could be considered reasonably by 2010.  The members also agreed that while the majority of intersection crashes involve either vehicles entering from perpendicular directions (the 'straight crossing path' scenario) or vehicles attempting to turn left ('left turn across path' scenario), that other types of intersection crashes, including collisions with pedestrians and bicycles, also should be addressed by intersection collision avoidance systems.

What About Stop Signs?

IC members also noted that candidate systems should be considered for addressing crashes at stop signs as well as traffic signals.

Countermeasures discussed at the workshop included relatively complex systems that could identify safety threats using traffic and pedestrian/bicycle sensors, analyze traffic movements in order to determine appropriate means to avoid the threat, and communicate the needed responses to drivers.  This communication could be either infrastructure-only or vehicle-highway cooperative.  Infrastructure-only systems might use innovative variable message/graphic signs or special strobe lights to communicate with all drivers;  for example, a strobe light could be used only when a driver is about to violate a stop sign to help the driver to recognize the presence of the stop sign.  Vehicle-highway cooperative systems would allow information to be processed and displayed right in the vehicle; for example, a message that warns of a potential collision with a traffic signal violator could allow the other driver to stop in time or otherwise to avoid colliding with the violating driver.

According to the minutes of the August workshop, participants also made the point that, in spite of the Secretary’s 'stretch goal' words regarding metropolitan areas, rural aspects should not be lost in the effort.  Federal officials indicated that USDOT is interested in developing countermeasures for both urban and rural intersections.

Significant New Area of Research

Although NHTSA performed a major study of intersection collision avoidance (ICA) systems over the past several years (see related articles in IVsource archives), this work is seen as merely a beginning.  FHWA states that the development of effective intersection collision avoidance systems will require a significant additional research effort.  At the meeting, the Consortium considered a research "strawperson" including tasks to: 

1. analyze crashes and mitigation concepts;

2. develop intersection collision avoidance concepts and algorithms;

3. develop analytical models to assess safety countermeasures;

4. develop or adopt infrastructure-based sensors;

5. examine human factors issues;

6. define vehicle-infrastructure (or vehicle-vehicle)  communication methods;

7. assess benefits, costs, and institutional barriers to deployment; and

8. develop in-vehicle systems. 

This group is not blind to the significant challenges involved in addressing ICA.  At the workshop, there was discussion of the type of causal factors that need to be examined, including roadway/intersection geometric factors such as sight-distance, grades, curves, etc.

Who's Got the Data?

There were also discussions regarding the understanding of human factors associated with intersection crashes, and the need to go beyond police reports to obtain crash data.  The key is "to know what is happening in the driver’s mind when approaching the intersection," according to the minutes.  It was suggested that naturalistic driving studies in the field and simulator research may be useful in understanding these human causal factors; naturalistic driving studies would provide data not only on crashes but near-crashes.  The possibility of the SAVME equipment to provide this type of data was discussed.  SAVME is a NHTSA-developed data collection system mounted on a tower overlooking intersections, which can collect information on vehicle trajectories and thus provide insight into near-crashes.

A dual strategy to approaching ICA was suggested:  a “bottom-up” analysis that uses existing crash data bases to examine rural/urban/suburban intersections, and a “top-down” analysis that examines driving performance at intersections using simulators and various analytical models.  Bob Ferlis of FHWA presented information on the use of crash causal factors to identify potential infrastructure-based countermeasures.  Examples were provided for straight crossing path and left turn across path scenarios at signalized and unsignalized intersections.  He also discussed the extension of these infrastructure-based systems to cooperative systems.

Jim Misener of PATH led a presentation from California, which outlined a proposal utilizing staff from PATH, Caltrans, and Visteon.   Jim indicated that team objectives included research, production, and deployment, with the overall goal of providing situational information that improves driver decision making at intersections.

Federal officials at the meeting indicated that USDOT is looking for one cooperative, unified program from the Infrastructure Consortium.  They asked the group to consider architecture issues, the applicability to signalized vs. unsignalized intersections, deployment issues, and the need for performance specifications.  A key issue revolved around who would be warned -- the driver violating the intersection controls or the driver in danger of being impacted?

Infrastructure-Only Systems for Starters

The group agreed that infrastructure-only systems would be deployed initially.  In the short term, infrastructure-only systems should be capable of evolving into infrastructure/vehicle communication systems. While the focus of the Intersection Collision Avoidance research should be on light vehicles, it might be possible to demonstrate on other platforms.  And, predictably, several familiar human factors questions came up: What is the impact on behavior?  Will these systems make drivers more aggressive?  What could be the unintended consequences?

Boosting the Effort: Standards, Contract Support, Programmatic Road Maps

Before long, the discussion turned to the need for standards.  While the initiation of standards development would have to begin soon if deployment is expected in ten years, considerable work still needs to be completed in the US even before agreement can be reached on the performance specifications that would lead to a standard by incorporation.  It was indicated at the meeting that standards work is ongoing at SAE and ISO and that this work may form a basis for intersection collision avoidance systems.  Participants noted the challenge of developing a standard for ten years in the future, given the pace of technology advances.

To help implement the research plan once it is finalized, a traffic engineering firm has been selected to work in this area, according to Ferlis, and will be starting work in October.  The firm will be able to work with and assist the states in analyzing crash databases.  A second firm may be brought on board by FHWA to define communication requirements.  The group felt that field operational tests and demonstrations should be part of the research plan, and the possibility of involving the Collision Avoidance Metrics Partnership (a GM-Ford research body) was raised.

After a full day of discussion, it was agreed that, in addition to the eight research areas listed above, tasks should also be added for an architecture study, a vehicle-vehicle communication study, and analysis of the integration of current Advanced Traffic Controllers used at signalized intersections.  Inclusion of development of algorithms and sensors for wet pavement conditions was suggested.   Further, there was agreement that promising concepts would need to be validated through field operational tests and that design and deployment guidelines would also be needed.  Ray Resendes, IVI Program Manager for USDOT, asked for a roadmap incorporating the research tasks discussed.

Joining state DOT representatives were technical experts from California PATH and Visteon.  Various members have raised the possibility of opening the group to accept private corporations as members, and this item is under active consideration.  But it appears to be a fact that the IC will work closely with the Enabling Research Consortium within IVI, which is composed of automotive companies.  There appear to be plans for both groups to meet in October to jointly define the initial stages of the research program for intersection collision avoidance.


[Top]

... contact Bob Ferlis, FHWA, at 202.493.3268 or Robert.Ferlis@fhwa.dot.gov.
 

[Top]