Collision Avoidance > Intersection Avoidance

Printer-friendly version

• Introduction
• The Rationale
• System Description
• Technologies Used
• Research 
• Implementation Challenges
• Implementation
• References

INTRODUCTION

Intersection Collision Avoidance Systems use sensors to gather information about vehicle movements near an intersection, process that information to determine if a collision is at risk of occurring, and issue warnings to drivers of vehicles in danger. They differ from traffic control signals in that they are continuously processing information when vehicles are present and creating messages tailored to specific vehicles' paths, speeds and driver behaviors.

There are three types:

• vehicle-based autonomous systems, which have been developed mainly by private sector manufacturers;
• infrastructure-based systems, where the warnings and sensors are located in roadside arrays;
• and systems that link vehicles to other vehicles or vehicles to infrastructure, so-called "cooperative" systems.

The latter two are more often the purview of state dots and government-supported research institutions because they would entail public investments.

Intersection crashes may be classified into one of the following five categories as shown in Figure 1:

a. Left Turn Across Path – Opposite Direction
b. Left Turn Across Path – Lateral Direction
c. Left Turn Into Path
d. Right Turn Into Path
e. Straight Crossing Path

Figure 1: Intersection Collision Scenarios

RATIONALE

Roughly 43 percent of vehicle crashes in the U.S. occur at intersections or are intersection-related. A significant share of them take place at intersections with traffic signals or stop signs. Their causes are often due to drivers' misjudgment of the situation, failure to correctly observe the situation, or inability to accurately perceive the degree of danger. Some 60% of rural intersection crashes occur even after the driver entering the main roadway has stopped before proceeding, researchers have found. Another significant share of crashes are caused by the driver entering the intersection against the signal or failing to stop at a stop sign.

These findings suggest that interventions such as warning systems and driver assistance could be particularly effective in reducing intersection crashes.

System Description

Vehicle-only systems consist of technologies developed for longitudinal and lateral collision avoidance, which can be divided into systems that:

• Advise or warn the driver (collision warning),
• Partially control the vehicle, either for steady-state or as an emergency intervention to avoid a collision (driver assistance); or
• Fully control the vehicle (vehicle automation).

Infrastructure based systems use enhanced traffic signals and no turn-warning, based on wireless communication and in-pavement sensors.

Infrastructure-vehicle cooperative systems integrate the above systems with in-vehicle hazard warning systems

technologies Used

A number of vehicle-only systems are similar to those used in longitudinal avoidance systems, including adaptive cruise control and forward collision warning.

Infrastructure-based systems detect traffic generally with radar or loop detectors, process a warning algorithm, then initiate dynamic signing.

Cooperative systems operate in a similar manner but communicate warnings internally to the vehicle’s internal hazard warning system via wireless networking. Warnings can be integrated with a driver interface.

Research

Intersection Decision Support

Research on infrastructure-based systems is currently being conducted under the Infrastructure Consortium, which is comprised of the U.S. DOT, and the California, Minnesota, and Virginia DOTs, which are sponsoring the Intersection Decision Support System research project. The research is being conducted by University of California at Berkeley PATH (Partners for Advanced Transit and Highways) Program, University of Minnesota Intelligent Transportation Systems Institute, and Virginia Polytechnic Institute/Virginia Tech Transportation Institute. While the consortium is primarily concerned with infrastructure systems, its teams are also doing limited research into collaborative systems.

Left Turn Across Path: Opposite Direction at UC Berkeley PATH

Left Turn Across Path: Opposite Direction crashes account for 27.3% of intersection related crashes in the US. Two-thirds of them occur at signalized intersections.

Reasons for these types of crashes include:

• Failure to judge safe gaps in traffic correctly
• Failure to judge speeds of closing vehicles correctly
• Obstruction of driver's view
• Failure to perceive opposing vehicle.

The PATH research is working with remote sensors for upstream traffic and loop detectors to detect downstream traffic and the subject vehicle. Sensors measure and relay range, rate and trajectories. The sensor data is processed in a warning algorithm. If a potential conflict is detected, a signal is sent to a dynamic warning sign, which will activate.

Left Turn Across Path: Lateral Direction at University of Minnesota ITS Institute

University of Minnesota’s ITS Institute demonstrated an IDS system that addressed Left Turn Across Path: Lateral Direction crashes, which typically occur in rural areas when a vehicle attempts to cross or turn onto a road at an unsignalized intersection. ITS Institute researchers found that 60 percent of crashes at rural intersections happen even after drivers stop before proceeding into the intersection. They termed this a gap perception problem.

The ITS Institute system was designed to tell a driver if it is unsafe to enter the main roadway. Radar detectors are deployed at five points around the roadway to detect approaching vehicles. The detectors communicate to a central processor via a wireless connection. The processor then runs an algorithm, which calculates which gaps are safe or unsafe to enter the roadway. Depending on the result, the algorithm may activate an LED no-left-turn signal.

Straight Crossing Path Crashes at Virginia Tech Transportation Institute

According to Virginia Tech Transportation Institute researchers, approximately 30% of intersection crashes involve vehicles executing a straight crossing path. The Institute research is focusing in ways to prevent those crashes caused by traffic signal and stop sign violations.

The system being tested includes pole-mounted radar at signalized intersections to determine an approaching vehicle’s speed and location and warn the driver with dynamic signing (LED stop sign and strobe light) if a violation is likely. An additional countermeasure that researchers have begun is a set of "intelligent" rumble strips that would deploy if a violation seemed imminent.

Vehicle-Based Systems in the Private Sector

Adaptive Cruise Control is now available in luxury passenger vehicles in the U.S. from Mercedes, Infiniti, and Lexus, as well as some trucks. These systems have been available for a number of years in a wide range of vehicles in Japan and Europe. Forward collision warnings, primarily from Eaton Vorad, have also been available and used on trucks for several years. Vehicle based systems are discussed in greater detail in the Longitudinal Avoidance report on this site.

Integrated Adaptive Cruise Control and Forward Collision Warning Systems: General Motors and Delphi Delco Electronics Systems

These companies are in the midst of a five-year, $35 million research project sponsored by the National Highway Traffic and Safety Administration (NHTSA) testing integrated Adaptive Cruse Control and Forward Collision Warning systems on Buick LeSabres. The system alerts the driver with sound and adjusts the cruise control speed if there is a possibility of a collision. This technology is now offered on select models of 2004 Cadillacs. University of Michigan Transportation Research Institute (UMTRI) is engaged in the Field Test of the system, which commenced in March 2003.

Implementation challenges

More research is needed on the causes of intersection crashes, the best technologies to detect potential crash scenarios, and the best warning systems to effect timely and appropriate responses from drivers or, in autonomous mode, from the vehicle's internal safety systems.

Additional hurdles are the development of interoperable systems, common performance standards and the costs of adoption.

Implementation

Intersection collision avoidance technologies are still in the research phase. In June 2003 a number of the collision avoidance technologies discussed here were demonstrated at the FHWA’s new test facility in McLean, Virginia in conjunction with the US DOT’s National Intelligent Vehicle Initiative meeting. Research on intersection decision support is slated to be complete by early to middle 2005.

References

"Crash-Prevention Technologies Showcased at the New FHWA Test Facility," ITS America. Vol. 13 No.7, July 2003.

Ferlis, RA. INFRASTRUCTURE COLLISION-AVOIDANCE CONCEPT FOR STRAIGHT-CROSSING-PATH CRASHES AT SIGNALIZED INTERSECTIONS. Transportation Research Record - Journal of the Transportation Research Board 1800. pp. 85-91. 2002

Frye, Cathy. "International Cooperation to Prevent Collisions at Intersections," Public Roads. July/August 2001.

Infrastructure Intersection Collision Avoidance: US DOT Intelligent Vehicle Initiative (IVI). FHWA. March 6, 2002.

"Inside the US DOT's Intelligent Intersection Test Facility," Newsletter of the ITS Cooperative Deployment Network. July 17, 2003.

Najm, W.G. and J. Koopmann, "Analysis of Crossing Path Collision Countermeasure Systems." Draft dated September, 2000.

Shladover, Steven. "Recognizing and Addressing Safety," ITS World. March April 2001."

Traffic Safety Facts 2002:A Compilation of Motor Vehicle Crash Data from the Fatality Analysis Reporting System (FARS) and the General Estimates System (GES)." US DOT and NHTSA. September 2003.

Wang, Hui et al., "Safer Roads Thanks to ITS," Public Roads. May/June 2002.

Author: Terri O'Connor. April 19, 2004