Driving forward cooperative intersection safety applications

Gregory Davis, Federal Highways Administration, John Harding, 834 National Highway Traffic Safety Administration, and Mike Schagrin, 781 ITS JPO (RITA) chart the course for cooperative intersection safety applications being pursued as part of the IntelliDrive programme.

Crashes at intersections accounted for 8,703 highway fatalities in the US in 2008. Research and development is moving forward on IntelliDriveSM safety applications designed to help drivers avoid intersection accidents. These new safety systems could substantially drive down the highway death and injury toll by using different combinations of infrastructure, vehicle and wireless technologies to gather, process and deliver information to the driver in real time.

Building on promising results from the Cooperative Intersection Collision Avoidance Systems (CICAS) programme, which has been folded under the umbrella of IntelliDriveSM, the 324 US Department of Transportation (USDOT) is planning additional research and development into four intersection safety applications: Violation Warning; Stop Sign Assist; Signalised Left Turn Assist, and Traffic Signal Adaptation.

Currently, the most advanced in development is the Violation Warning prototype application, which is ready for a full-scale Field Operational Test (FOT). Researchers are continuing to conduct and analyse pilot validation tests of the Stop Sign Assist working prototype, which is projected to be a cost-effective alternative to installing traffic signals at rural intersections. Engineering tests of a prototype Signalised Left Turn Assist application are planned for next year. The Traffic Signal Adaptation application is still in the early testing stages.

Here is a synopsis of research findings so far, and the next steps:

Violation Warning

The Violation Warning system uses vehicle-based calculations and infrastructure input to offer drivers an in-vehicle warning when another vehicle commits a possible traffic signal or stop sign violation - potentially averting the dangerous scenario of one driver being unaware that another is 'rolling through' or running a stop sign or traffic signal.

Thus far, this system has shown great promise in initial tests of its effectiveness, which included a 2007-2008 pilot for an FOT with 87 drivers at the 5593 Virginia Tech Transportation Institute.

Researchers concluded that the prototype Violation Warning System was capable of correctly identifying an impending violation of both traffic signals and stop signs, and subsequently deliver a timely alert to warn the driver. The system performed reliably and did not deliver any false negative alert. In addition, the infrastructure equipment performed well and required virtually no maintenance.

Overall, the system performed better than the performance specifications required.

The next step for this application is a test in real-world conditions, so an engineering test to determine system limits was conducted as a part of the pilot. During this test, a Violation Warning-equipped vehicle drove behind a tractor-trailer that blocked DSRC messages from the intersection. The system performed well, despite the interference, and the results of the pilot indicated that the Violation Warning application is ready for a real-world FOT.

The FOT, which will involve many more drivers over a longer period, will be designed to study the potential benefits of the Violation Warning system; to learn more about customer acceptance factors, and to reveal any unexpected safety consequences of the prototype system.

Stop Sign Assist

Many of us have personally experienced that frightening moment when we underestimate how much time is needed to safely make a left turn from a complete stop onto a major road with speeding traffic. In fact, misjudging the gap in traffic needed for a safe turn is responsible for about two-thirds of crashes on sections of rural roads that intersect high-speed divided highways. While adding a traffic signal is traditionally considered the best approach to reducing crashes at such locations, these signals often create a new set of safety and operational problems-including increased rates of rear-end vehicle collisions and lower levels of service on the major road.

An innovative solution to this problem is the Stop Sign Assist application, which uses infrastructure-based sensors and dynamic message signs to help drivers make better decisions on when to proceed onto or across a high-speed road.

This application not only helps drivers make safer decisions, but offers a lower-cost solution for transportation authorities operating on tight budgets and timelines. A Stop Sign Assist application prototype is projected to cost less than $200,000, which is slightly less than the cost of a rural signalised intersection retrofit.

Beyond the actual cost, the necessary technology and component devices are available now. The sign is based on a typical, commercially available display that uses off-the-shelf colour LED components, making the sign readily available for state and local agency procurement.

The Stop Sign Assist application has the additional benefit of not stopping traffic along the mainline, leading to less congestion and other disruptive incidents that require local resources.

The Stop Sign Assist application was developed through a study led by the 2103 Minnesota Department of Transportation and the 584 University of Minnesota's ITS Institute, with nine state DOTs participating in its development. Of the nine, eight of the states also participated in studies to determine human factors that would affect the design of the system to ensure that any proposed concept would offer the highest benefit towards driver safety.

Participating researchers subsequently developed a prototype sensor system. In 2008, this project evaluated driver interface design concepts by having 60 users test the system using driving simulation equipment. The field-based prototype system, which consists of the sensor suite and the selected driver interface design concept,, was pilot-tested at a Minnesota test intersection in the fall and winter of 2008.

Results of the Minnesota pilot tests showed that the Stop Sign Assist warning sign is easily understood, and has a small but positive effect on driver behaviour. The pilot demonstrated that drivers chose to cross through safer (greater) gaps in traffic when using the Stop Sign Assist sign. The Stop Sign Assist sign also appears to have increased the frequency of two-stage manoeuvres - where the driver stops in the median after crossing the first leg of the artery to assess the gap in traffic approaching from the right. Drivers reported that they found the sign easy to use and beneficial.The bottom line is that 80 per cent of crossings were two-stage with the Stop Sign Assist sign turned on compared to 60 per cent when the sign was not operating, indicating that drivers can significantly benefit from this technology, and in turn, make everyone's driving experience safer.

A similar field test of the Stop Sign Assist application is currently being conducted in Wisconsin.

Under the IntelliDrive programme, a larger-scale, three-year FOT is being planned which will take place at three test intersections (one in Minnesota and the others yet to be determined). Although the system is effective as a solely infrastructure-based solution, the added benefits of vehicle connectivity will also be assessed.

Signalised Left Turn Assist

For transportation professionals who are focused on improving safety, the following scenario is a familiar one: a car in the left turn lane at a busy intersection has a green light, but must yield to oncoming traffic. As oncoming traffic speeds past, the driver creeps forward in anticipation, waiting for the right moment to turn. The driver sees a gap in the traffic, and begins to turn but misjudges the speed of an oncoming tractor-trailer. The driver of the tractor-trailer slams on the brakes and hits the horn. The turning car just barely makes it; another near miss.

Left turns at intersections are particularly dangerous, and more than a quarter of signalised intersection crashes occur when drivers make a left turn after passing through a permissive green light.

The Signalised Left Turn Assist application in development by the California Partners for Advanced Transit and Highways (California PATH) has the potential to help drivers avoid making unsafe decisions in these situations and in turn, save lives and prevents serious injuries. This application uses both in-vehicle and infrastructure-based driver warning displays to alert drivers when they are about to make an unsafe left turn.

Past research on left turn behaviour has revealed that drivers making left-turn decisions can be influenced to choose a safer gap in traffic when using both in-vehicle and infrastructure-based warning displays. Both types of warning device were found to have the potential to reduce unsafe turns caused by poor judgment of gaps, or the inability to see oncoming traffic.

Researchers have learned that most drivers make their initial turning decision about 2.5-3 seconds before the turn. Most drivers who participated in the research study preferred to have the alert 3-4 seconds in advance of the turn. The Signalised Left Turn Assist system improves driver decision-making, and intersection safety, by giving them more time to make a better decision.

The system works by combining roadside sensors, infrastructure-based messaging signs, communications and positioning technologies, dynamic maps, traffic signal interfaces, and Dedicated Short-Range Communication (DSRC). It is the most complex of the IntelliDrive intersection safety applications, because it addresses multiple interacting vehicles and other road users (such as the presence of pedestrians) within a difficult and dynamically changing environment. It is this complexity which requires use of a wide range of technologies.

Engineering tests to gauge system performance of the Signalised Left Turn Assist application are planned for 2009 at a California intersection in the Palo Alto area. These tests will also evaluate driver acceptance of this system, and establish a baseline of data for comparison in a full-scale field operational test.

Traffic Signal Adaptation

If there is one element that distinguishes the act of running a red light, it is the element of surprise. When a driver commits a red light violation, whether through being distracted or by attempting to 'beat the yellow', other drivers have to act quickly to prevent a dangerous situation from becoming a deadly one.

Developed by California PATH researchers, the Traffic Signal Adaptation application is designed to reduce red-light-running crashes by changing the traffic signal to an all-red status when one is imminent, helping to quickly alert potential victims of the situation before it is too late.

The Traffic Signal Adaptation application operates by combining the infrastructure-sensing features of the Signalised Left Turn Assist with the vehicle-based violation warning features of the Violation Warning application.

Using vehicle-generated data, the Traffic Signal Adaptation system identifies the probability of a red light violation and transmits the data to infrastructure systems: traffic detectors act in coordination to provide further data on location and speed of other vehicles and to gather real-time traffic signal phase timing status from the signal controller.

The system then sends a message to the signal controller to adjust the phasing to an all-red status, thus preventing the potential victim from entering the intersection.

While this system is currently in the engineering testing phase, research results to date have shown that the Traffic Signal Adaptation application has the potential to reduce right-angle crashes by over 55 per cent. This research also revealed an important insight: only 10 per cent of red light running occurrences are caused by vehicles being caught in the 'dilemma zone' (the point when a yellow light causes the driver to decide whether to run the yellow light or to brake suddenly and risk a rear-end crash).

These results highlight the importance of developing the ability of the Traffic Signal Adaptation system to correctly predict when a true red light running incident is about to occur and adjust signal timing accordingly. The extra second or two that this system provides could be the difference between a close-call and a deadly accident.

V2I safety roadmap 

The IntelliDrive intersection safety applications research is part of USDOT's IntelliDrive Vehicle to Infrastructure (V2I) research programme. The V2I programme is focused on the prevention of crash types and scenarios that are associated with high fatality and injury rates, including intersection crashes, roadway departure crashes, speed-related crashes, and commercial/transit vehicle crashes.

The technologies under investigation will provide a graduated spectrum of safety interfaces, including in-vehicle information and advisories; and in-vehicle driver warnings of imminent crash scenarios.

These interfaces will be based on open standards for data and communications, including DSRC for time-critical applications; and an open platform concept for non time-critical applications (standards that will require full engagement with stakeholders to provide a successful framework for the full deployment of these applications across the country).

To promote a collaborate development of this framework, the USDOT is planning to work with stakeholders to create a V2I Safety Roadmap. This document will chart the course of these safety technologies, as the Department and the transportation community work towards moving these life-saving applications from the research phase to real-world deployment.