How Human Drivers Can Benefit From Collective Perception: A User Study

Abstract

Collective perception is one of the key ideas of vehicular networking and allows the exchange of data about perceived objects. However, unlike autonomous driving systems, human drivers cannot screen large numbers of objects to judge their dangerousness. An assistance system in the vehicle, therefore, must do this job. This article shows a concept for a human–machine interface that could be used to warn the driver in case such a system detects an actually dangerous object. A user study in a driving simulator was performed to evaluate its potential to prevent accidents. Eye-tracking glasses were used to analyze the driver’s gaze during different types of situations. Furthermore, the participants’ subjective experience was evaluated with a questionnaire. Results show that drivers trust the system and brake earlier and with more control due to the warnings, and ultimately, the majority of accidents could be avoided thanks to the warnings.