Research In Action

Like everyone, I was deeply saddened to learn about the first reported fatality attributed to a vehicle in auto-pilot mode. The death of Joshua Brown, who crashed in Florida as his vehicle was supposedly in auto-pilot mode might be remembered as the first self-driving motor vehicle fatality.

While the technology may still sound like science fiction to most of us, hearing about this fatality brings into question the reliability of self-driving technology. One single incident should not however cause us to rush to conclusions. On the day Mr. Brown died while his car was in auto-pilot mode, an estimated 100 other fatalities occurred on US roads. How many of those crashes could have been prevented and fatalities averted if all vehicles were equipped with such technology?   

Auto-pilot mode, which is a close cousin to autonomous and self-driving technology, is still in its infancy but has a lot to offer in terms of safety. In the US, every year over 30,000 people lose their lives in car crashes. Teen drivers who have a lot less experience, crash 3 to 4 times more than adults. On average, seven teens die every day on US roads. As a researcher who studies how to improve the safety of children, adolescents, and young adults, the question I ask is: Can we harness this technology to decrease teen crash risk or do we need to be concerned about unintended negative consequences for novice teen drivers?

Traffic safety researchers at the Center for Injury Research and Prevention are responding to this challenge with projects like the Simulated Driving Assessment led by Dr. Flaura Winston, which used novel simulator-based technology to assess teen drivers’ performance in crash scenarios, and a more recent naturalistic driving project led by Thomas Seacrist and me. This project is analyzing data from 3,000 fully-instrumented vehicles to assess self-driving challenges and risks.

As self-driving vehicles start populating our roads, the expectation outlined by NHTSA is that “the driver is still responsible for monitoring the roadway and safe operation and is expected to be available for control at all times and on short notice. The system can relinquish control with no advance warning and the driver must be ready to control the vehicle safely.” Understanding differences between this recommendation and how actual drivers behave is critical as we consider the overall safety benefits of this innovative technology. Researchers at the Center for Injury Research and Prevention are working to address these important questions.

The Significance of Human Factors

While manufacturers race to produce the first fully autonomous car, little attention has been paid to the human factors at play for teens when using the technology. Should newly-licensed drivers with little behind-the-wheel experience be trusted with the technology? Will they be more inclined to use their phone and text than more mature drivers if they feel that the vehicle is “driving for them”? How do age and experience factor in?

A team of CHOP researchers, led by Dr. Aditya Belwadi and me, have made it a priority to understand the human factors at play in self-driving and auto-pilot technology so as to inform future NHTSA recommendations.

One study funded through the Center for Child Injury Prevention Studies utilizes a driving simulator to immerse drivers in self-driving technology. In this study, participants will first be asked to demonstrate their driving skills. The simulated vehicle will then be put in self-driving mode and after some time, the driver will be exposed to two scenarios with auto-pilot failure that will require an immediate take-over and a strong evasive maneuver.

With this study, the researchers hope to:

• Assess the ability of teen and adult drivers to sustain attention for long periods of time while in self-driving mode
• Measure how quickly drivers react and avoid an imminent crash and stabilize the vehicle
• Assess driver’s impressions of the self-driving mode before and after experiencing it

The stakes are enormous. NHTSA estimates that motor vehicle crash deaths rose 8 percent between 2014 and 2015, registering the highest year to year increase in 50 years. As the economy improved in 2015, crash fatalities went up: 38,300 people died on the road in 2015.

As more and more autonomous vehicles appear on US roadways, it is critical that we fully understand how drivers will interact with and use the technology and then set up adequate safeguards, perhaps federal regulations, to ensure improved protection in this new reality.