Threat Assessment Design for Driver Assistance System at Intersections

The field of road safety and safe driving has witnessed rapid advances due to improvements in sensing and computation technologies. Active safety features like anti-lock braking systems and adaptive cruise control have been widely deployed in automobiles to reduce road accidents. However, the US Department of Transportation (DOT) still classifies road safety as “a serious and national public health issue.” In 2008, road accidents in the US caused 37,261 fatalities and about 2.35 million injuries. A particularly challenging driving task is negotiating a traffic intersection safely; an estimated 45 percent of injury crashes and 22 percent of roadway fatalities in the US are intersection related. A main contributing factor in these accidents is the driver’s inability to correctly assess and/or observe the danger involved in such situations.

This project focuses on assisting human drivers with negotiating busy intersections in the presence of possibly errant drivers with uncertain intentions. We are developing a novel design for a threat assessment module (TAM), which combines a learning-based intention predictor with an efficient sampling-based threat assessor to compute the threats of errant drivers in real-time. This threat data is used to evaluate the safety of several possible escape paths, which may be proposed to the human driver if evasive maneuvers are warranted. The approach is demonstrated through experimental results in the RAVEN facilities.

The picture below shows a human-driven (middle-front), and autonomous (acting as possible errant) vehicles near an intersection in the road network of RAVEN. The human-driven vehicle receives warnings from TAM when an errant vehicle detected in its vicinity is predicted to collide with it.