ABSTRACT
The design of advanced driver assistance systems faces the challenge of providing automated support behavior that complements the human driver safe and efficient. In order to explore and evaluate the coordination between man and machine, human emulation as a simulation alternative is already acknowledged (cf. Wizard of Oz approach). However, validation for such an approach within the context of designing driver support, is currently missing. This paper reports a validation study concerning the use of human emulation during the design of advanced driver assistance. For this, an automated and an emulated version of a lateral support system were compared in a fixed-base driving simulator setup. Participants received a directional precue on the steering wheel and used this information to choose the safe direction in an upcoming time-critical situation in order to avoid a rear end collision. Differences in precue onset were observed for the automated and emulated version, but the distinct characteristics in terms of timing did not reveal an effect on driver reaction times. Present results not only reveal a relatively high tolerance for precueing onset, they also suggest human emulation as a valid simulation alternative during the design process of driver support.
