Crowdsourced Traffic Event Reporting: A Driving Simulator Study

Abstract Safe and efficient road transportation today bases on reliable real-time data about various traffic events and conditions. Despite reliable infrastructure-originated data, crowdsensing promise comprehensive coverage of traffic events. However, in general, we cannot consider crowdsourced traffic reports as objective, but must take into account also the reporter’s perception of the events. Even similar events can be perceived differently, depending on the location, time, current traffic situation, and peoples’ current state and previous experience. We assessed drivers reporting preferences in an experiment performed in driving simulator, while drivers’ behavior was observed using professional eye-tracking device. Additionally, their preferences were assessed before the driving experiment using a questionnaire as well as after the experiment in semi-structured interviews. Results show that there are significant delays between the time of event observation and reporting. Moreover, several wrong reports, multiple reports with the aim to correct them, and also behavior that lead to lower driving safety was observed.