abstract background pattern

Open Source: When Will Autonomous Vehicles Save Lives?

Compiled by Jackie Powder

Automated vehicle technologies already are saving lives. Adaptive cruise control, automatic emergency braking and lane-keeping assist are in many cars and trucks out on the road. And manufacturers testing fully automated vehicles already have video evidence of cars avoiding crashes with pedestrians and cyclists that human drivers would be hard-pressed to replicate. As we develop rules for automated vehicles, an important task is figuring out how to measure the benefits. If an autonomous vehicle slams its brakes, leading the car behind to tap its bumper, should that be counted as a crash? Or is that a life saved because the autonomous vehicle didn’t hit the child who darted in front of it?

Mark R. Rosekind, PhD, is the administrator of the National Highway Traffic Safety Administration.

Vehicles with various levels of automation are already saving lives. Adaptive cruise control (available since 2001) and electronic stability control (available since 2012) are technological advances that have made cars safer. The promise of fully autonomous vehicles includes enhanced mobility for many individuals as well as fewer crashes, injuries and deaths. Some experts expect autonomous vehicles to save 30,000 lives annually. Predictions are challenging because of many unknowns. How, where, and when will these vehicles be widely available? How will they interact with the existing vehicle fleet? And, how will the public respond? Based on the success of seat belts and airbags (estimated to have saved more than 600,000 lives from 1960–2012), the potential of autonomous vehicles is great. 

Andrea Gielen, ScD '89, ScM '79, is director of the Johns Hopkins Center for Injury Research and Policy.

This is a difficult question and perhaps a tipping point in transportation safety as we know it. The answer may vary based upon how one views the engineering developments associated with automation, and the very definition of what is an autonomous vehicle. If one considers automated systems such as active emergency braking in today’s production vehicles as features that have been developed in our pursuit of self-driving or highly autonomous vehicles, one might say life savings have already occurred. If one is looking to fully autonomous vehicles with no driver involvement, life savings may be decades away.

Bryan Reimer, PhD, is a research scientist at MIT.

The question implies that vehicles are autonomous or not, when it is actually a continuum. Autonomous features in vehicles are already preventing crashes and saving lives. As an example, forward-collision warning with automatic emergency braking (autonomous braking) is a proven, life-saving technology that will soon be standard in most new cars. Autonomous features (and one day, autonomous vehicles) won’t prevent every crash—and for the foreseeable future, drivers will need to maintain their skills for when autonomous features fail. But they will keep us safe from many of the things which place us at risk on the road today.

Johnathon Ehsani, PhD, is an assistant professor in Health Policy and Management.

Current automotive safety technology saves lives, but it is difficult to measure the effectiveness of current technologies without gathering information about when they help a driver avoid an accident. Autonomous vehicles will probably begin being sold in the early 2020s, but won’t become a large part of the fleet until around 2030. If they are involved in a fatal accident, the sensory data gathered by its 360-degree view will allow a more objective and measurable re-creation of the accident. But until near-miss data is captured (a Big Data project) we will not have a good estimate of the total safety value of autonomous vehicles.

Bruce M. Belzowski, MA, is managing director of the Automotive Futures group at the University of Michigan Transportation Research Institute.