Is the automated driving system “aware” of the pedestrian? Results from two driving simulators
Mon-B17-Talk I-01
Presented by: Vanessa Stange
In urban areas, drivers often have difficulty detecting pedestrians who want to cross the road but are obstructed by parked vehicles. Highly automated vehicles (HAVs, SAE Level 4) need to solve this problem independently and at least as safely as human drivers. We conducted two experimental studies, a stationary driving simulator study and a human&vehicle-in-the-loop study, to investigate whether passenger trust in automation can be increased and perceived risk in the driving situation can be mitigated by adjusting the driving behaviour of the HAV. We varied the lateral offset to the centre of the lane (left, right, none) and deceleration (only in the driving simulator study) when passing a parked vehicle on an urban main road (speed limit set at 50 km/h) to convey to the passenger that the vehicle is “aware” of a potential hazard caused by the presence of the parked vehicle and the pedestrian hidden by the parked vehicle. The results support the idea that it is beneficial for human passengers to be informed of the HAV's “awareness”. While the deceleration somewhat enhanced passengers’ trust, the introduction of a lateral offset did not have a substantially positive effect on passengers' perceived risk or trust. Future research should compare the effects of behavioural adjustments with other approaches, such as in-vehicle visual warnings and information systems.
Keywords: Traffic psychology, automated driving, passenger, urban area