A multi-modal architecture for intelligent decision making in cars

Abstract

This paper describes a software architecture named "Gatos" engineered for intelligent decision making. The architecture is built on a distributed multi-agent system cougaar. Gatos provides a solution for sensor fusion. We propose using multiple sensors to monitor driver status, driving performance, and the driving environment in order to address bad driving behavior. Our approach for a Driving Monitor is based on both monitoring and regulating driver behavior. The system is designed to intervene and to interact with the driver in real time (if possible) to regulate their behavior and promote safe driving. A prototype is implemented using a driving simulator, but infrastructure buildup and new in-vehicle technologies make this a feasible solution for vehicles on the road. © Springer-Verlag Berlin Heidelberg 2007.