Abstract
This paper proposes a methodology to detect low attention state based on an individual driver reference model in car-following state. First, a driver-vehicle reference model in car-following state, treating the host vehicle and the preceding vehicle as a 1-DOF spring-mass-damper system, is used to express the normal driving of the driver. Next, low attention driving state is detected by comparing the current driving data with the value estimated by the reference model. Finally, the validity of the proposed objective evaluation is investigated by comparing to subjective evaluation of drivers and video image of the driver’s facial expression during driving on highway.
Keywords: Driver state detection, Driver modeling, Naturalistic driving study, Statistical modeling, Car-following