Whilst in theory autonomous vehicles will operate without any human intervention, in the short-term it is likely that they will need to be designed so that humans can regain control over the vehicle when required.
How that occurs however is still up for debate. I wrote last year about a Stanford study that analyzed the ease with which drivers can regain control, and the results suggest it is far from straightforward. The study placed a number of test drivers on a track to find out, and the results suggest the transition isn’t as smooth as we might think.
“Many people have been doing research on paying attention and situation awareness. That’s very important,” the authors say. “But, in addition, there is this physical change and we need to acknowledge that people’s performance might not be at its peak if they haven’t actively been participating in the driving.”
Taking back control
This was also the topic of a second study, this time from the University of Southampton, which examined just how long it might take for control to go back and forth between man and machine.
The study saw a number of participants in a driving simulator, where they were driving at 70mph. Sometimes they were otherwise engaged in separate tasks, other times they were not. The researchers wanted to test the response time when they both ceded control, and regained it from an automated system, with intervention requests sent at random intervals during their drive.
Perhaps not surprisingly, whenever the participants were engaged in something other than focusing on the road, their response time was much longer. Even then though, the response time was anything from uniform, with some responding in just 1.9 seconds, but others taking up to 25.7 seconds. This represents a huge challenge for designers when building devices that cater for such a wide range.
The researchers hope that their paper will help to guide policy makers when they establish guidelines for the next generation of autonomous vehicles. It is especially crucial, they note, that policy makers are aware of the differences inherent in response times under critical and noncritical conditions.
“If normal, noncritical control transitions are designed based on the data obtained in studies utilizing critical situations, there is a risk of unwanted consequences,” they urge.
For instance, if if we legislate for too short a lead time, it could prevent drivers from responding optimally, causing dangerous maneuvers, whether it’s sudden braking, accidental swerves and so on. These kind of actions could be justified in emergency situations but incredibly dangerous in normal scenarios.
It underlines some of the complexities inherent in our transformation to a driverless world, and it will be fascinating to observe how manufacturers and policy makers alike tackle it.