Traffic conflict
A traffic conflict, in transportation engineering, is an event involving two or more moving vehicles approaching each other in a traffic flow situation in such a way that a traffic collision would ensue unless at least one of the vehicles performs an emergency maneuver. Traffic conflicts are defined by their time-to-collision, post-encroachment-time, and angle of conflict parameters as well as the vehicles' position in time and space.
Traffic conflicts have typically been used for transportation safety studies, whereby observing and monitoring individual collisions may be impractical, unfeasible, or unsafe. Traffic conflicts are used as traffic collision surrogates, under the assumption that the same factors affecting collision rates also affect conflict rates, in proportion to the conflict severity, termed conflict hierarchy.
The principles of traffic conflicts apply to all modes of transportation involving vehicles operating in a non-guided medium, including motorized vehicles, airplanes, boats, and bicycles.
By modes of transportation
Motorized vehicles
Most traffic conflicts involving motorized vehicles are observed on highways, usually involving lane changing or sudden changes in vehicle speeds (rear-end collisions), or in intersections, involving a large array of conflict types.
Methodology
Traffic conflicts are complicated events to define visually. Observers are trained to spot traffic conflicts using the same set of videos.
Recently, research has shifted towards automated conflict detection with the advent of traffic video analytics and faster processing power. The Federal Highway Administration-sponsored Safety Surrogate Assessment Model is a tool being developed for the analysis of conflicts using simulated results from traffic microsimulation.
Criticism
Some research suggests that the correlation between conflicts and collisions is weak, or can vary from site to site.
Because human observation of conflicts is a complicated process, some argue that a certain amount of subjectivity invalidates traffic conflict analysis involving human observers.