Driver Impairment Diagnostic to improve Reliability and Acceptance of Automotive Driver Assistance Systems - Application of fuzzy expert methods

Abstract

Vehicle and passenger safety is a critical issue for public authorities as well as for car manufacturers and suppliers. Besides drivers training and information many technological advances have been performed since a couple of years and have drastically improved the road safety. Nevertheless the core problem remains the driver itself. Several studies held both in US and in Europe show that 90% of the accidents are due to intentional or non-intentional behaviors of the drivers, for example a bad perception or knowledge of the environment but also reduced physiological or psychological conditions. The Vehicle/Driver/Environment process can easily be compared to a closed loop system where the driver plays the role of the management/control unit. The driver pilots the vehicle considering the information it delivers, its perception of the environment and according to the targets and planning that have been set. Furthermore, the drivers behavior is influenced by its driving aptitude and capacity and should also be disturbed by environmental conditions. Despite the development of more and more sophisticated Advanced Driver Assistance Systems (ADAS) that substitute the driver in critical situations, it is currently not possible to fully exclude it from the vehicle management loop. Furthermore, the deployment of ADAS will really be efficient and accepted since it will fit with drivers needs, aptitudes and capacities to deal with complex environmental contexts, but also when providing an adapted assistance through a human centered design. Thus, an important issue is to get a better on-line knowledge about the driver and its limitations. Within this context, original functions providing an on-line driver state diagnostic about sleepiness and visual or cognitive distraction have been studied. These works are including the design and the development of complete systems from the measurement settings, up to the Human Machine Interaction concepts. A special effort was dedicated to design reliable drivers diagnostic processes. As the managed information is often imprecise or uncertain the use of qualitative reasoning methods has been foreseen. Finally, these functions and systems have been successfully implemented on several experimental vehicles and evaluated in real driving conditions.