Realtime Technologies

News

Human Factors Research: Using Sims to Increase Comfort and Reduce Confusion with “Autonomous” Vehicles

No conversation about autonomous vehicles (AV) and advanced driver-assistance systems (ADAS) gets far without celebrating its potential for decreasing human misery. Estimates of traffic fatality reductions in the neighborhood of 75–80% are not uncommon (see, for example, Liu, P., Yang, R., and Xu, Z., “How safe is safe enough for self-driving vehicles?”, 2019, in the journal Risk Analysis).

But if we hope to see any of the promised benefits of AV/ADAS—let alone these huge safety gains—we need to broadly adopt these technologies. Unfortunately, according to annual surveys taken by AAA, two-thirds to three-quarters of Americans consistently describe themselves as “too afraid to ride in a self-driving vehicle.”

Based on survey responses and interviews, AAA believes that a good deal of this can be attributed to fear of the unknown. But they also note that “confusion about the capabilities of present day ADAS equipped cars also could be contributing to the uncertainty. AAA found that 40 percent of Americans expect partially automated driving systems—with names like Autopilot, ProPILOT or Pilot Assist—to have the ability to drive the car by itself.” (These systems emphatically do not have such capabilities.)

But there is hope. AAA also found that as drivers interact with AV and ADAS-equipped vehicles, they become significantly more comfortable with them. “On average, drivers who [own vehicles with ADAS technology] are about 68 percent more likely to trust these features than drivers who don’t have them.”

That’s promising! But it’s also a bit of a chicken-and-egg problem: people embrace AV/ADAS once they get some experience with AV/ADAS, but they are unlikely to buy/experience a vehicle with ADAS until they become comfortable with it.

This is an interesting and important (if inherently challenging) intersection where “driving research” meets “human factors research” meets “civil engineering” meets “public policy.” How can we help motorists experience AV/ADAS driving if they are unwilling to get in an AV/ADAS vehicle and give it a whirl?

Human Factors Research Shows Sim Experience Can Increase Acceptance of New Driving Technologies

In recent years, many universities have been using their immersive driving simulators to explore how people will truly interact with AV/ADAS-equipped vehicles (as opposed to how they believe they will interact with and feel about those technologies).

Dr. Joseph F. Coughlin, Ph.D. (founder and director of MIT’s AgeLab) has often referred to their Realtime Technologies (RTI) immersive simulator as a “lie detector on wheels”—a way to get at the truth of how drivers actually behave behind the wheel. MIT has also repeatedly confirmed that “behavior patterns measured in the [RTI] driving simulator correspond with driving in the real world driving and driver self-report.”

RTI’s sims have proven uniquely suited to such human factors research because of the existing SimADAS/SimDriver plug-ins, combined with the SimObserver plug-in (which allows for easy recording and synchronization of study participant behavior recorded multiple independent sources, including audio, video, in-simulation behaviors, and third-party hardware).

Recent studies have had promising results in how we might break free of the “chicken-and-egg AV/ADAS experience” snare.

Sherrilene Classen and her colleagues at the University of Florida recently used their RTI simulator to introduce older adults to using AV/ADAS. They found that, with older adults, a simulated AV exposure was as effective as riding in an actual autonomous vehicle in terms of improving their perception of AV/ADAS as safe. In general, real-life or simulated demonstrations of AV/ADAS boosted confidence in the technology more than a passive experience (e.g., a demonstration video or written material).