Can the Startle Reflex be Manipulated to Reduce Take-over Time in Pre-crash Scenarios for Autonomous Driving?

Principal Investigator: Valentina Graci, PhD, Children’s Hospital of Philadelphia

WHAT WAS THE PURPOSE OF THIS PROJECT?

CantheStartleReflexbeManipulated

Drivers of autonomous vehicles may be slow to take over vehicle control in time to avoid a crash. We wanted to understand if a novel take-over warning based on the startle reflex could accelerate take-over reaction times in adults and newly licensed teens. We also wanted to see how age and engagement in the driving task (ready to react versus texting while driving) influence the acoustic warning’s effectiveness. We placed seven adult drivers and seven teenage drivers on a lateral oscillating sled to undergo a simulation of pre-crash swerving events preceded by the acoustic startling warning (acoustic startling pre-stimulus, ASPS). They were initially seated with their hands on their laps and instructed to grab the steering wheel when the sled began to move.

WHAT DID YOU FIND AND WERE ANY OF THE RESULTS SURPRISING?

We used computational modeling to create full-vehicle and CRS models and used a combination of Q6 ATDs and human body models to simulate the occupant. This is one of the first projects in which we used a pediatric human body model called the PIPER model, which was developed specifically for injury prediction in crash tests.

WHAT WERE YOUR FINDINGS?

We found that when adult drivers were ready to react, they lifted their hands from their lap towards the steering wheel more quickly and reduced their trunk and head lateral motion when exposed to the ASPS. In contrast, when exposed to the ASPS, the teens did not act faster and showed more lateral displacement than did the adults. We were surprised to find that our acoustic startling warning can “startle” adult drivers into a more advantageous position within the seat belt, in addition to accelerating hand motion toward the steering wheel.

WHAT ARE THE INDUSTRY IMPLICATIONS FOR THIS RESEARCH?

Our research can help industry to develop more effective take-over warnings. Since the ASPS activates later than forward collision warnings and lane departure warnings, it could be used in conjunction with those systems as a last resort if the driver does not respond to them. Our findings can be used to inform ADAS for all vehicles.

WHAT’S NEXT FOR THIS LINE OF RESEARCH?

Our first study only involved male drivers, and we want to find out if the startle reflex could be beneficial to female drivers as well. We also want to understand if the differences we observed with age – where adult drivers’ movements were accelerated by the ASPS but those of teens were not – apply to females as well.

Project Team Member

Jalaj Maheshwari, MS, Children’s Hospital of Philadelphia

Students

Rahul Akkem, Drexel University; Madeline Griffith, University of Pennsylvania; Catherine Krawiec, Rochester Institute of Technology; Hanh Do Phung, Drexel University

IAB Mentors

Jonathan Gondek, Calspan Corporation; Mike Kulig, Calspan Corporation; Emily Thomas, Consumer Reports; Mark LaPlante, Graco Children’s Products Inc.; Jerry Wang, Humanetics Innovative Solutions Inc.; Suzanne Johansson, General Motors Holdings LLC; Julie Kleinert, Technical Advisor