Empowering ADAS With Driver-Supervised Learning Of Preferences: Parameterization And Human-Machine Interaction

The development and widespread use of advanced driver assistance systems (ADAS) has made considerable progress; however, in order for drivers to be confident in delegating driving tasks to these systems, there is a need to further improve their acceptance and trust. One approach to achieve this goal is to provide the driving system with the ability to autonomously learn a driver's preferences under the driver's supervision through a human-machine interface (HMI). This requires first of all identifying these driving preferences, which are derived from the driving situation to be managed, and from the driver's behavior. This paper presents the first results of an experiment conducted on a full-scale simulator in order to validate the parameterization of driving preferences and to identify future interactions through the HMI between drivers and the proposed new driving assistance system.