Communicating Driver Intents: A Layered Architecture for Cooperative Active Safety Applications

Abstract

The great promise of vehicle to vehicle communications includes a reduction or even elimination of collisions and fatalities on roadways, especially of those due to driver error. A major roadblock to the effectiveness of these systems is the market penetration of cooperative Driver Assistance Systems. Many proposed and existing implementations of cooperative systems are only effective if a majority of vehicles are capable of cooperating. We propose the use of a Layered Architecture for Cooperative Active Safety Applications (LACASA) that includes driver behavior or intent prediction as a fundamental building block for cooperative assistance systems. Simulation results show how such information could improve safety even in limited-deployment scenarios, and in wide-scale deployment, unsafe maneuvers are nearly eliminated.