Driver Perception of Lateral Collision Threats 2020-01-1198
Immediate collision hazards pose obvious threats to approaching drivers and therefore provoke emergency evasive responses. When the hazard is a vehicle intruding into the lane ahead, how its movement characteristics influence an approaching driver’s response is not well understood.
This study examined the relationship between intruding vehicle motion and hazard perception. Seventeen subjects viewed first-person perspective recordings of a simulated vehicle travelling down a two-lane roadway containing several intersections with stop-controlled minor roads. Stopped vehicles were located at approximately half of the minor road intersections. Throughout the study, some vehicles (termed ‘intruders’) accelerated into the subject’s lane of travel at 1 of 6 pre-determined acceleration rates. Subjects were instructed to ‘brake’ their vehicle by pressing the space bar on a keyboard as soon as they perceived that a collision was imminent. Perception and response times (PRT) were measured as the elapsed time between the intruder’s first motion and braking. The duration of the perception interval was subsequently estimated by deducting a simple reaction time interval from the total PRT.
The results indicated a significant relationship between perception time and intruder vehicle motion. Subjects took longer to perceive slower accelerating intruders as immediate hazards, and therefore responded when intruders were closer to the subject’s path of travel, compared to intruders accelerating at higher rates. Differences in perception time across conditions of intruder acceleration suggest that the start of the intruder’s motion was not the sole source of information used in detecting an immediate hazard.
To the authors’ knowledge, this study is the first to quantify the relationship between acceleration levels of laterally intruding vehicles and hazard perception processes. Such findings aid forensic assessments of driver behaviour during emergency situations and in designing autonomous vehicle systems that account for human limitations.