New research has found that drivers often fail to spot hazards missed by automated vehicles, and that this failure only gets worse the longer drivers spend riding in them.
The study – Driver Vigilance in Automated Vehicles: Effects of Demands on Hazard Detection Performance – was conducted by researchers from Rice University and Texas Tech University, who examined the behaviour of 60 licensed drivers operating an automated car in a simulator.
"The bottom line is, until automated driving systems are completely reliable and can respond in all situations, the driver must stay alert and be prepared to take over."
To mimic the effects of automated driving, those taking part were told they would not need to operate the simulator’s steering wheel, brake pedal or accelerator pedal.
They were then instructed to monitor the roadway for vehicles stopped dangerously at intersections and intruding into the driver’s lane – a hazard the automated vehicles could not detect. As part of this monitoring, participants had to distinguish between safely stopped vehicles and those that were stopped dangerously at intersections.
During the first 10 minutes of the simulation, the drivers’ success rate was, at best, close to 88%, found the authors, suggesting that all the drivers missed some hazards. Over the course of the 40 minute simulation, the drivers’ accuracy dropped between 7% and 21%.
One possible reason for this is that people get used to cars doing the driving and become complacent, said study co-author Pat DeLucia, a professor of psychological sciences at Rice. DeLucia also pointed to previous research showing that people are terrible at monitoring for hazards that occur infrequently.
“The bottom line is, until automated driving systems are completely reliable and can respond in all situations, the driver must stay alert and be prepared to take over,” said Eric Greenlee, an assistant professor of psychological sciences at Texas Tech and the study’s lead author. “And this research clearly shows that is not happening, and gets worse as time passes.”
“These vehicles have a lot to offer, but we’re a long way from being able to detect everything going on,” wrote the researchers.
“Until that day comes, we hope this research will raise awareness about the limitations of automated cars and their operators.”