OPTIMIZED SYSTEMS

Traffic Model Uses Human Reason to Understand Crashes

Rebecca Kutzer-Rice | Date: 09-30-11 | Comments

Researchers in Israel have created a traffic-modeling tool that predicts how drivers and pedestrians will react to surrounding events. The simulator could enable city planners and engineers to design safer roads and fix regulation problems.

A new tool from researchers at the University of Tel Aviv combines robotics and statistics to analyze driver and pedestrian behavior. Called SAFEPED, the computer model could become an invaluable tool for traffic planners and engineers.

Traffic �black spots� occur when certain intersections experience unusually high numbers of accidents. As the causes of accidents in such places are often unclear, city planners lack effective tools to prevent them from occurring. The most realistic traffic model ever, SAFEPED promises to help minimize the occurrence of dangerous intersections.

SAFEPED could help researchers identify causes of fatal accidents and prevent future crashes before they occur.

SAFEPED analyzes each individual car and person as a separate entity. Unlike previous models, which do not take into account human reason, SAFEPED is able to predict how each moving �agent� will react to surrounding events.

�Because drivers and pedestrians behave according to the same habits and rules at any intersection they approach, we presumed that the problem lay in the environment.� Gennady Waizman, a PhD student at Tel Aviv University�s Geosimulation Lab at the Department of Geography and the Human Environment, explained in a statement. �With this program, we can model a real intersection in the simulator, and make modifications to the environment or traffic regulations to see how they impact the safety of the junction.�

Researchers programmed agent behavior by analyzing video footage and traffic statistics from the Israeli police force. They also spent hours observing how pedestrians and drivers reacted to each other.

SAFEPED allows researchers to view crashes from the perspectives of every involved agent. This enables them to determine whether visibility issues arose or if a pedestrian or driver made a bad decision. Researchers can then change variables like traffic regulations and road design to prevent the virtual accident.

In addition to Waizman, researchers included Itzhak Benenson and Shraga Shoval, both professors at Tel Aviv University, as well as M.A. student Eilon Blank-Baron. The team has presented SAFEPED at conferences in Jersualem and London.

More information, as well as a video of SAFEPED in action, can be found here.