Context-Aware Driving Assistance: An Approach for Monitoring-Based Modeling and Self-learning Cars

Recent cars are equipped with a large number of sensors, electronic and communication devices that collect heterogeneous information about the vehicle, the environment and the driver. The use of the information coming from all these devices can highly contribute to the improvement of the vehicle safety as well as the driving experience. The last few years were marked by the development of a large number of in-vehicle intelligent systems that use driving behavior models to assist the driver ubiquitously. However, an important aspect to enhance driving experience is to make the provided assistance as close as possible to the behavior of the car owner, hence a need of personal models of drivers learned from their observed behavior. In this paper, the concept of intelligent and self-learning car is presented and examples of some car's embedded systems are given. Also, the role of modeling driver behavior in the design of driving assistance systems is emphasized. Further-more, the importance of monitoring-based driving behavior model construction to enable a personalized assistance is brought out together with some potential applications of formal driving behavior models.