Waymo Explains How It Keeps Test Drivers Attentive & Safe From Fatigue

Energy
Wednesday, August 31st, 2022 11:30 am EDT

In the past, I’ve written several times about the attention problem with advanced driver assist systems. The fundamental problem is that when you aren’t having to steer the vehicle, it’s hard to keep yourself alert and awake, and a number of studies bear this out. Waymo sums it up like this:

“… a general relationship between an increasing degree of automation and a corresponding reduction in human performance exists. This general phenomenon has been termed the “irony of automation”. In other words, and in the context of driving, the better the automation, the less attention human drivers will pay to traffic and the system, and the less capable they will be to resume control.”

But there aren’t hundreds or thousands of accidents piling up with systems like Tesla Autopilot, despite the number of people using such systems on the road.

We have to keep in mind that systems like Autopilot are pretty mature. Are they perfect? Obviously not. But, they aren’t making frequent catastrophic mistakes, either. So, the attention problem doesn’t happen at the same time as a system error to generate a huge number of accidents.

What about people who are testing less mature systems? It turns out that Waymo has thought about this a lot as it put its test drivers on the road with software that can make a lot of mistakes at the worst times.

In a recent paper Waymo released, the company spills the beans on its methods and strategies for avoiding problems with driver fatigue and inattention. While this information might not be useful for most drivers, hopefully the people testing other systems (including Tesla’s FSD Beta and Autopilot) can get a few good ideas from what they shared.

Waymo says its Fatigue Risk Management (FRM) framework revolves around three things: (1) fatigue prevention; (2) fatigue monitoring; and (3) fatigue mitigation. On a practical level, this involves (1) continued education; (2) awareness and reporting; (3) real-time vigilance assessment; (4) supplemental engagement; and (5) adaptive scheduling. Let’s take a quick look at what each of these things mean.

Continued Education

Waymo starts its effort with education. If its test drivers understand the problem, and how to deal with it (the other four blocks), they’ll be able to avoid fatigue-related accidents.

Waymo says specialists need continuous, comprehensive instruction for preventing, spotting, and handling fatigue so they’re always aware of risks related to fatigue. The best approach begins with how to prevent fatigue using tried-and-true methods (like self-care and good habits) before moving on to how to educate oneself about the signs of tiredness and what measures can be taken both combat them in the moment and mitigate future risk.

This idea applies far outside of Waymo’s operations. Many owners of vehicles with ADAS systems either aren’t aware that a fatigue problem exists, or are in denial when people try to say that they could be falling asleep while using systems like Autopilot or FSD Beta. Being aware of the risk and knowing more (below) about how to avoid the risk, is the first step in staying safe.

Awareness & Reporting

To be more aware of their own fatigue, Waymo’s test drivers have to fill out Periodic Fatigue Surveys (PFS). The PFS, which is an online survey accessed by specialists on their own devices, is a practical way for them to report fatigue to supervisors and get help in the short- and long-term. It’s not meant to be used by employers as a performance measurement tool, though. In each PFS survey, specialists give a subjective self-assessment of their level of fatigue or alertness.

This gives drivers an opportunity to report to their supervisors their actual situation, and this allows supervisors to take action to keep things safe. It also adds accountability for both drivers and supervisors who might be tempted to do what dispatchers do to truck drivers (force them to keep going). Actions in response to high fatigue include taking breaks, prescribing a period of physical activity, and other things that generally work well in preventing tiredness and inattention.

Real-Time Vigilance Assessment

Waymo uses Driver Monitoring Systems (DMS) in much the same way many ADAS systems, but with more monitoring and communication. An automated fatigue monitoring system may provide multi-modal notifications to the autonomous specialist, which might include a tone, a vibration under their seat, and a flashing light. It can also instantly provide remote human judges with a real-time video feed.

Supplemental Engagement

The framework we’ve covered above include comprehensive fatigue prevention and monitoring. But, Waymo says they’re not enough on their own. They say active intervention measures should be included in a comprehensive FRM, according to data analysis of over 10 years of on-road operations. Supplemental engagement with the driver also helps keep people alert and awake. This comes in two forms: interactive cognitive tasks (to keep the mind awake) and secondary alerts.

They don’t go into detail on what a cognitive task is, but they go into great detail describing how the tasks are assigned. During monotonous driving tasks, both the driver and an automated computer program can assign these little tasks to keep people awake. One description of the tasks sounds like the driver is given a quiz or asked questions about the weather and road conditions, which they can answer by pressing a button on the steering wheel.

This sounds a lot like suggestions we all heard in Driver’s Ed telling us to sing along with music or do something else to avoid fatigue, but much fancier and timed based on studies.

Secondary alerts for manual driving are related to the vehicle’s control transition from automated to manual, which follows a primary alert triggered at the time of control transition. A secondary alert is used to minimize the chance of an inadvertent control transition from automated to manual (for example, when fatigue or attention is distracted).

Adaptive Scheduling

Finally, Waymo doesn’t follow strict schedules that could exacerbate fatigue. Instead, they’ve both read the relevant academic literature and done their own studies to determine how to schedule driving time.

When it comes to shifts, they’re careful to give good breaks, time off, and schedule shifts in such a way as to minimize fatigue. They are also careful to give good breaks to avoid fatigue from piling up during a shift. Finally, they aren’t afraid to give a fatigued test driver other work if they just aren’t safe to drive that day. Waymo has plenty of other work to do, and puts drivers on that work to keep them and the public safe.

We Could All Learn From These Techniques

Waymo’s study and implementation of these techniques is a lot more complicated than I presented here. If you’re curious, I’d recommend reading the whole paper. Anybody driving a lot, and especially people testing things like FSD Beta could learn a thing or two about staying awake and alert.

Featured image provided by Waymo.

 

Appreciate CleanTechnica’s originality and cleantech news coverage? Consider becoming a CleanTechnica Member, Supporter, Technician, or Ambassador — or a patron on Patreon.

 


Don’t want to miss a cleantech story? Sign up for daily news updates from CleanTechnica on email. Or follow us on Google News!

 


Have a tip for CleanTechnica, want to advertise, or want to suggest a guest for our CleanTech Talk podcast? Contact us here.


Advertisement

 


This post has been syndicated from a third-party source. View the original article here.