Human Factors Research: Building Trust in ADAS and Autonomous Driving
Published
Advanced driver assistance systems (ADAS) and automated vehicles (AV) have the potential to save tens of thousands of lives and prevent millions of auto accidents each year. This isn’t just a talking point from someone selling “self-driving” cars. According to multiple analyses, various ADAS features (like adaptive cruise control, automated emergency braking, and lane assist) are already successfully preventing accidents. The Insurance Institute for Highway Safety (IIHS) and the Highway Loss Data Institute (HLDI) has found that just two popular ADAS features—forward collision warning and automated emergency braking—when taken together, prevent 56% of front-to-end crashes resulting in injury. In other words, those two features, by themselves, can save more than 6,000 lives and prevent 250,000+ injuries each year.
But drivers are turning off the ADAS systems at an alarming rate. According to the J.D. Power 2019 U.S. Tech Experience Index (TXI) Study, 23% of owners with ADAS features like lane assist found the alerts annoying, and 61% of those annoyed motorists subsequently disabled them. This discontent even informs purchasing decisions. J.D Powers found that more than 90% of drivers who didn’t have ADAS features wanted them on their next vehicle—but only about 60% of those who’d had them wanted them again.
“Automakers are spending lots of money on advanced technology development,” explains Kristin Kolodge, executive director of Driver Interaction and Human Machine Interface Research at J.D. Power, “but the constant alerts can confuse and frustrate drivers. … It’s essential that the industry recognize the importance of an owner’s first experience with these lower-level automated technologies because this will help determine the future of adoption of fully automated vehicles.”
“What We Have Here Is a Failure to Communicate”
According to David Bouwkamp, Executive Director of Business Development at FAAC, “The issue is communication from the systems to the occupants allowing the occupants to have confidence in those systems. When an occupant disengages these systems for these sorts of reasons, it is a failure to communicate. The occupants have lost confidence because they do not believe or do not understand what the system is trying to do or tell them to do.”
This particular observation—that failed communication leads to confusion, annoyance, and a critical lack of confidence in ADAS—aligns with some of the research already being done using RTI simulators. For example, in 2018, researchers collaborating in Australia and Michigan published their findings on the relationship between how easy an ADAS feature is to use (that is, how clearly the ADAS communicated) and drivers’ trust in that automated driving system during a simulated experience using an SAE Level 3 automated vehicle in a situation that required several periods of transfer-of-control between the ADAS and the human occupant.
These researchers reported that, “Consistent with previous research … the trust [users] described improved as a result of the experience in the driving simulator…. Such work similarly suggests that there may be critical interactions between ease-of-use and trust, as noted [elsewhere]. … Participants were typically describing ease of use in a way that reflected their confidence.”
Crafting Custom Software and Simulation Solutions
For decades, FAAC has been an industry leader in building simulation solutions for training and research. Their Realtime Technologies division (RTI) regularly helps researchers, institutions, and OEMs use custom simulators to run real-time human-in-the-loop simulation, modeling, and research. RTI sims include:
Extensions for adding 3rd party systems and studying driver interaction with these features
Scenario development tools that make it easy to trigger ADAS systems and messages (vital to any usability studies)
In-cab display GUI customization options (important for crafting effective ADAS-to-driver communication strategies)
These features have proven especially helpful for usability studies exploring how to effectively notify drivers with ADAS warnings and status.
This particular study was completed using a fixed-based RTI simulator built around the cab of a Nissan Versa and equipped with immersive audio, visuals, and haptics. But more important than the simulator’s hardware was the software.
A Platform for Quickly Developing ADAS Usability Studies
Realtime simulators use the latest version of RTIs SimCreator DX scenario authoring and development software. With SimCreator DX, researchers and engineers can quickly develop and refine complex, custom multi-body driving scenarios without writing a line of code.
SimCreator DX can now be expanded with several modules especially developed for rapid prototyping and exploration of autonomous and ADAS solutions. The SimADAS and SimDriver software modules make it easy for researchers to rapidly iterate through different approaches, exploring human-in-the-loop interactions with various levels of automation, including SAE level 1 and 2 driving automation, as well as transfer-of-vehicle-control situations in both city and freeway driving environments. SimCreator and its modules come ready to use, with a full library of pre-installed generic maneuvers, vehicle dynamics, maps, and modules. That said, every element is fully customizable, and can be readily adapted to match OEM specifications.
According to some estimates more than 90% of motor vehicle accidents are attributable to human error. ADAS is a promising tool for directly addressing roadway safety. But without further research into driver trust and implementation, many of those tools won’t be adopted. If you would like to discuss simulation options to study human behavior factors that affect ADAS efficacy, please reach out and let’s talk about simulation solutions that can achieve your research goals.