A multiagent based warning system for pedestrian safety

A pedestrian warning system is often used in detecting the movements and intentions of a pedestrian and giving warnings to the relevant participants. Studies in pedestrian-vehicle conflicts have focused on modeling the behaviors concerning the participants, developing the sensors installed on the vehicle, giving warnings to the driver, etc. In this study, in addition to give the drivers warnings, the warning system uses traffic signals to take the initiative to prevent the pedestrians from crossing a street. The information concerning the degree of risk is given to the driver for making decisions along with a braking system that is triggered in emergency. We have developed an agent-based simulation system for a pedestrian warning system that has a traffic light and an active warning system on vehicles to achieve driver personalization and improve road safety.