Human-machine cooperation principles to support driving automation systems design

Car manufacturers and automotive suppliers design more and more Advanced Driving Assistance Systems (ADAS) that are quickly installed into cars. Nevertheless, such ADAS address different types of driving activities and driver's behaviors; they might concern longitudinal and lateral controls, lane changes, navigation, and try to take into account driver state and behavior. But what about the complementarity of such different ADAS if they are all together implemented in one car? And what about their interactions with the driver? Interactions have to be defined according to the levels of automation selected by the car manufacturer. In the French national project CoCoVeA (French acronym for Cooperation between Driver and Automated Vehicle), three levels of automation have been studied involving several existing ADAS. And this paper proposes a methodological approach to identify competences and capacities of driver and ADAS in the case of many driving situations in order to check complementarity, function allocation, authority management, as well as cooperative aspects in order to assess reliability of such a Human-machine system and the acceptance and even the attractiveness of highly automated vehicle.