Implicit communication in cyclist-vehicle interaction: Examining the influence of driving dynamics in interactions with turning (automated) vehicles on cyclists' perceived safety, behavioral intention, and risk anticipation

Abstract

Successful and safe management of interactions between cyclists and motorized vehicles often includes implicit communication, such as vehicle motion signals. While the introduction of automated vehicles (AVs) is anticipated to increase safety for cyclists by eliminating human error, it is still unclear how they should communicate implicitly in complex traffic scenarios. The turning maneuver constitutes a complex and ambiguous scenario, as it involves vehicle deceleration even when there is no intention to halt for the cyclist, proceeding straight ahead. A video-based laboratory study incorporating a bicycle setup examined the influence of four driving dynamics and automation status on cyclists' risk anticipation, perceived safety, and intention to continue to cycle through an intersection. Forty-two participants viewed videos of an oncoming vehicle making a left turn at an intersection from the perspective of a cyclist traveling straight through the intersection. The results indicated that participants perceived early and continuous braking of the vehicle ("passive yield") as the most favorable approach (reduced risk anticipation, higher perceived safety, and greater intention to continue to cycle through the intersection), even when compared to later but stronger braking ("active yield"). The objectively riskiest maneuver that would result in a collision ("collision") was evaluated as the least desirable. A rule-violating yet objectively safe maneuver (accelerating to exit the conflict area before the cyclist; "no yield") led to increased risk anticipation but not to lower perceived safety or reduced intention to continue to cycle through the intersection compared to active yield. The displayed automation status had minimal influence overall, although participants in the passive yield condition expressed greater uncertainty regarding their intention to continue to cycle through the intersection in interactions with AVs compared to MVs. It appears that natural driving dynamics exhibited by "good" human drivers could represent a promising approach to ensuring the safety and comfort of cyclists in interactions with AVs. Here, early and continuous braking is preferable to later and stronger braking.