Share this article:
Slow Responses and Safety in Semi-Automated Vehicles
Autonomous vehicles are becoming more popular – they are being developed by car manufacturers and technology companies, and they’re being tested on the road in a lot of markets. It’s disruptive innovation – it will fundamentally change a lot about the transportation domain. In the research community, we’re trying to understand how you design and implement autonomous vehicles and how they should function with the driver.
In the future, a driver will not have to use the steering wheel to control their car’s lane position or use the brake or accelerator pedals to control its speed; they may be much more likely to engage in more active secondary tasks like watching a movie, which is visually intensive.
But what happens if something goes wrong in the car or if there’s an emergency that means the driver has to take over? We need to fully understand the implications in terms of safety and how people will respond to this technology.
In our new study in the Journal of Safety Research, we used a driving simulator to explore this behavior. Before we see widespread use of self-driving cars, we’re likely to be driving semi-automated vehicles. So we asked participants to come in and either manually drive a car, controlling it with the pedals and steering wheel, or drive a semi-automated car, in which they did not need to control lane keeping or speed.
As we know, when using technology, things are designed to work well but they don’t always. So when something stops functioning, how does the driver respond and safely become the operator of the vehicle? To investigate this kind of situation, we simulated a gust of wind that pushed the cars out of their lanes and then monitored the drivers’ responses.
In the manual cars, the drivers had their hands on the wheel so they could steer back into the lane quickly and smoothly. But when we gave the drivers of the semi-automated cars something else to concentrate on – a video – they responded more slowly and severely, pulling the steering wheel harder and jolting the car back into the lane to compensate for the delay.
It may be tempting to say it is too dangerous for drivers to engage in non-driving tasks while controlling an automated vehicle, but in the future this kind of behavior may be the norm. And people are going to be distracted whether or not they are doing something illegal.
Ultimately, we want to improve people’s safety. There’s great opportunity in the semi-automated driving system because there are such huge safety implications: it’s life or death, not just for the driver or passenger but for anyone in the environment.
If we are to improve safety, we first have to understand how the technology is shaping drivers’ behavior – then we can shape the technology around that behavior.
For example, we can think about where in the vehicle a driver is carrying out a task (like watching a video) and how we might design alarms or alerts to help bring the driver’s attention back to the driving task. Many of our colleagues are working on how to design alarms and alerts to give drivers feedback about their performance and help them direct their attention to the appropriate task.
We are now continuing this line of research, looking at how we implement some alarms, when we do that, and how effective those types of alarms are. We are focusing on semi-autonomous systems and identifying some of the design features that might impact their safety.
When we think about safety it’s important to think about driver performance and how we assess that. This has implications for driver education: how do we shape or implement education for driving vehicles in which people are not actively engaged in some of the driving tasks? Will we have driving students learn how to do specific things even if the vehicle is designed to do the tasks for them?
Think about the transition from manual to automatic gear shift – someone who doesn’t know how to drive a manual transmission vehicle is still able to drive and knows the rules of the road. Take that concept and extrapolate it: compare a manually driven car to a vehicle in which the ‘driver’ has never driven. For fully automated vehicles, should there be driver education at all? I think so. But what that will look like in the future is still an important question.
About the Authors
David M. Neyens, PhD MPH is an assistant professor of Industrial Engineering at Clemson University. He received his PhD from the University of Iowa in 2010 and a Masters of Public Health in 2008 also from the University of Iowa.
Sijun Shen, PhD is currently a research scientist at the Center of Injury Research and Policy at the Nationwide Children’s Hospital. I received my PhD from Clemson University in 2016. He also earned a MS in Mathematical Science Department at Clemson University and a MISE in Industrial and Systems Engineering at Auburn University.
In addition to our acclaimed journal collection in Transportation, did you know that we recently launched a portfolio of books as well? For more information, click here.
The general scope of social sciences is vast, and Elsevier’s collection of journals, books, and eBooks examine in detail a wide range of topics in this area, from sociology, law, and cognitive science to political science, education, and linguistics. Our Chandos imprint in particular, known for high-quality scholarship in Asian studies, library and information science, and business management, reflects Elsevier’s continuing commitment to these crucial areas of study.