Drivers' comfort zone boundaries when overtaking pedestrians - Analysis of naturalistic driving and field test data

Abstract

In the European Union, while pedestrian fatalities decreased over the past decade, their relative percentage in respect to all traffic fatalities slightly increased. Worldwide, the share of the pedestrian problem, in road accident injury, has not shown any significant change or improvement in the last few years. Furthermore, considering the constantly increasing tendency towards vehicle automation, it is even more important to understand, quantify, and model road users comfort zone boundaries. Comfort zone boundaries turn out to be relevant regions for enabling automated vehicles to drive in such a way that users feel both safe and comfortable. Considering that safety research has thus far been focused on the interaction between drivers and pedestrian in the intersection scenario, the aim of this project has been to provide information about relevant safety metrics in the longitudinal scenario, i.e. when pedestrians are overtaken by a vehicle. Thus, the work described in this thesis involved quantifying the comfort zone boundaries when drivers overtake a pedestrian, both from the driver and the pedestrian's perspective. Driver comfort zone boundaries were estimated from naturalistic data in the UDRIVE database. As well drivers' behaviour was observed from a pedestrian viewpoint using a wearable data logger, which has been set up and employed in an experiment performed on the road. The analysis of 68 events from the UDRIVE naturalistic driving study and 481 overtaking manoeuvres from the road experiment has allowed to estimate the comfort zone boundaries, by mean of the application of Bayesian regression models. Driver behaviour was shown to be in uenced by the presence of oncoming vehicles, as well as by the walking direction of the pedestrian. Also, the steering input for the last phase of the overtaking manoeuvre, was surprisingly performed by drivers before the vehicle had passed the pedestrian. Overtaking maneuvers are complicated events, resulting usually in outcomes that cannot easily be predicted and are sometimes completely unexpected. However, this study showed how specific factors in uence drivers' comfort zone boundaries while overtaking a pedestrian. The two data sets investigated resulted to be complementary, allowing to have a complete understanding of the event. Findings are relevant for active system development, and Euro NCAP safety features assessment related to the collision pedestrian longitudinal adult (CPLA) scenario. Safety factors have been compared with previous study related to bicycle overtaking events.