Information of Waymo’s L4 and Tesla's L2 AV Systems – How to Design a Safer Road System with AV Technology
Waymo

Information of Waymo’s L4 and Tesla's L2 AV Systems – How to Design a Safer Road System with AV Technology

"It's likely in 20 years that we'll look back on those days when we gave a 17-year-old kid with less than 20 hours of experience control of a 5,000-lb. vehicle speeding along at 65 MPH as a somewhat insane practice." - Peter Diamandis 

Both human drivers (especially new or drunk, drowsy, doing drugs or distracted driver human drivers) and AV systems need to be monitored and controlled to make sure they drive properly and within the limits of their ability.

Question: would you rather have a human or AV driving the vehicle if you are walking or biking?

Answer: depends if the human is paying attention and cares about the life of the walker or biker or if the AV company designs their systems properly and drivers use the system as designed.

Humans drivers - 33% of all US traffic deaths are from striking a pedestrian/cyclist – in Maricopa Az, it is 41%

“Waymo Public Road Safety Performance Data Waymo pulls back the curtain on 6.1 million miles of self-driving car data – There were 47 contact events that occurred over this time period, consisting of 18 actual and 29 simulated contact events, none of which would be expected to result in severe or life-threatening injuries.”

Waymo pulls back the curtain on 6.1 million miles of self-driving car data in Phoenix - https://www.theverge.com/2020/10/30/21538999/waymo-self-driving-car-data-miles-crashes-phoenix-google

“I think it will go a long way to force other automated driving companies to reveal these kinds of data moving forward,” he said, “so when things go wrong, they provide a framework of data that is available to the public.”  In its first report on its autonomous vehicle operations in Phoenix, Arizona, Waymo said that it was involved in 18 crashes and 29 near-miss collisions during 2019 and the first nine months of 2020. These crashes included rear-enders, vehicle swipes, and even one incident when a Waymo vehicle was T-boned at an intersection by another car at nearly 40 mph. The company said that no one was seriously injured and “nearly all” of the collisions were the fault of the other driver.

https://blog.waymo.com/2020/10/revealing-our-approach-to-safety.html

 https://waymo.com/safety

See the report https://storage.googleapis.com/sdc-prod/v1/safety-report/Waymo-Public-Road-Safety-Performance-Data.pdf

In total, the Waymo vehicle was involved in 20 events involving contact with another object and experienced 27 disengagements that resulted in contact in post-disengagement simulation, for a total of 47 events (actual and simulated). In two of the actual events (which occurred after disengagement), post-disengage simulation revealed that the event would have been moderately more severe had the trained operator not disengaged. Therefore, these two events are treated in this paper according to their more severe simulated outcomes, yielding a total of 18 actual outcomes and 29 simulated outcomes.

No alt text provided for this image

Information on Tesla – note, each quarter, Tesla’s AV is getting better:

“In the 3rd quarter 2020, [Tesla] registered one accident for every 4.59 million miles driven in which drivers had Autopilot engaged. For those driving without Autopilot but with our active safety features, we registered one accident for every 2.42 million miles driven. For those driving without Autopilot and without our active safety features, we registered one accident for every 1.79 million miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 479,000 miles.” https://www.tesla.com/VehicleSafetyReport

The good news is: if the Waymo L4 and L2 Tesla systems are used correctly, you are much safer than in a car with a human driver without this technology. 

The bad news, If you use L2 system as if it was L5, you are a much more dangerous driver.

See the NHTSA’s Automated Driving Systems (ADSs – SAE International Automation Levels 3-5) Voluntary Safety Self-Assessment (VSSA) Disclosure Index.

https://www.nhtsa.gov/automated-driving-systems/voluntary-safety-self-assessment

Here is some information on past data from Waymo

JAN 6, 2020 - Waymo’s autonomous cars have driven 20 million miles on public roads

https://venturebeat.com/2020/01/06/waymos-autonomous-cars-have-driven-20-million-miles-on-public-roads/

2016 “Waymo has logged over two million miles on U.S. streets (2016) and has only had fault in one accident, making its cars by far the lowest at-fault rate of any driver class on the road— about 10 times lower than our safest demographic of human drivers (60–69 year-olds) and 40 times lower than new drivers.” Allison Crow

see this podcast: 205: Why Tesla's "Full Self-Driving" Beta Is Dangerous https://www.autonocast.com/blog/2020/10/29/205-why-teslas-full-self-driving-beta-is-dangerous

I would note, if not used properly. If drivers do not use the technology, people will die but if drivers do not use the technology properly, people will die as well.

From 2019, AAA report

https://newsroom.aaa.com/2019/10/aaa-warns-pedestrian-detection-systems-dont-work-when-needed-most/

Overall, the systems performed best in the instance of the adult crossing in front of a vehicle traveling at 20 mph during the day. In this case, the systems avoided a collision 40% of the time. But, at the higher speed of 30 mph, most systems failed to avoid a collision with the simulated pedestrian target. The other scenarios proved to be more challenging for the systems:

adult crossing in front of a vehicle traveling at 20 mph during the day. In this case, the systems avoided a collision 40% of the time.

When encountering a child darting from between two cars, with the vehicle traveling at 20 mph, a collision occurred 89% of the time with current AV technology. At night, none of the systems detected or reacted to the adult pedestrian. [night is when you need the system to work the most]

This demonstrates how important it is to educate the public that AV technology systems have limited safety benefit right now and how it is so important to NOT relay on this technology to allow for driver distraction. 

As different ADAS technology is deployed in a car, as the driver of the car, it is important to remember that until we reach level 5 AV, the driver is the captain of the car. To paraphrase William Ernest Henley poem Invictus: 

“It matters not how strait the road,

    How charged with ADAS technology the vehicle,

You are the master of your fate,

    You the captain of your car”

The challenge for AV technology Safety is the long tail problem. Billions of miles need to be driven to prove which technology is safer, as highlighted by the following statement: The Senate Committee on Environment and Public Works how autonomous vehicles will affect infrastructure, cybersecurity and public safety had the following quote:

  • “In 1910, there were 45 deaths for every 100 million miles traveled. … about one death per 100 million miles traveled today thanks to Federal motor vehicle safety standards and investments in safer roadways.
  • “The average driver has one police reportable crash approximately every 18 years of driving. The safest 10% of drivers— meaning those who remain alert, attentive, and sober, [do not speed] behind the wheel—are 10 times safer than the average driver. Many of the safest drivers will go through a lifetime without experiencing a single serious crash. Thus, to achieve the same safety as the best drivers, automated vehicles can almost never crash.” - Dr. Zachary Doerzaph
  • In 2016, on average a person was killed in a traffic collision every 84.7 million miles traveled on U.S. roads. Before the fatal crash in Arizona, Uber had reportedly logged 2 million autonomous miles as of the end of 2017 and was predicted to accrue another 1 million miles over the next 100 days. 
  • Based on a simple evaluation of this data, the autonomous Uber had one fatality in three million miles; that is a fatality rate 28 times that of human drivers.” Shaun Kildare 

https://www.epw.senate.gov/public/index.cfm/hearings?ID=FA6787B5-F941-4C2A-983D-2DE781560503

My view: AV/ADAS technology does not compensate for the drunk, drowsy, doing drugs or distracted driver. It is likely that it will take over 5 year for AV/ADAS technology to be more reliable than a human driver who is paying attention and cares about the lives around them.   All it takes is 1 distraction to cause an accident.

See the following report:

Cadillac's Super Cruise Outperforms Other Driving Assistance Systems - Other automakers close in on Tesla's Autopilot, now a distant second, in Consumer Reports' new ratings of 17 systems - By Mike Monticello October 28, 2020 https://www.consumerreports.org/car-safety/cadillac-super-cruise-outperforms-other-active-driving-assistance-systems/?fbclid=IwAR2gwu3E4tEx4pP10X60S2AkxuCoFpn4vP2nVe8fMXKigqSopszh6fRPgwc#ranking

Pedal-powered and electric bicycles (e-bikes), electric scooters (e-scooters) and other small transportation devices – rose to 136 million in 2019, a 60% increase over the previous year. - 40 percent of the injured individuals were first-time e-scooter riders https://www.ghsa.org/resources/understanding-and-tackling-micromobility-transportations-new-disruptor 

 The IIHS study also notes that nearly 40 percent of the injured individuals were first-time e-scooter riders, pointing to inexperience as a factor in crashes. This, too, tracks with GHSA’s recommendation to increase on-device training. Learn-to-ride events coupled with ongoing practice can help individuals build skill and reduce their crash risk.

The number of trips on shared micromobility devices – pedal-powered and electric bicycles (e-bikes), electric scooters (e-scooters) and other small transportation devices – rose to 136 million in 2019, a 60 percent increase over the previous year. As trips have increased, so too has the potential for crashes, with hospitals reporting triple digit spikes in e-scooter injuries and hospital admissions.

The sad fact is that technology is likely not to fix the traffic accidents any time soon with Drug and Alcohol prevalence 64.7% testing positive for at least one active drug in Seriously and Fatally Injured Road Users. Drug and Alcohol Prevalence in Seriously and Fatally Injured Road Users Before and During the COVID-19 Public Health Emergency.

Drug prevalence during the public health emergency with 64.7% testing positive for at least one active drug compared to 50.8% before the public health emergency began.

Drivers also showed an increase in testing positive for two or more categories of drugs going from 17.6% before the pandemic to 25.3% during.

Of particular note, active THC was more prevalent among drivers during the public health emergency than alcohol (32.7% versus 28.3%), and opioid use among drivers almost doubled going from 7.5% to 13.9%.

The U.S. Department of Transportation’s National Highway Traffic Safety Administration today released a preview of 2019 data from the Fatality Analysis Reporting System and preliminary estimates for the first half of 2020. Alongside the release of the 2019 preview data and 2020 first-half fatality projections, the agency issued two companion reports, Special Report: Examination of the Traffic Safety Environment During the Second Quarter of 2020 and Drug and Alcohol Prevalence in Seriously and Fatally Injured Road Users Before and During the COVID-19 Public Health Emergency.

https://www.nhtsa.gov/press-releases/2019-fatality-data-traffic-deaths-2020-q2-projections

Doug Hohulin

To Save 1 Billion Lives with AI, Exponential Blueprint Consulting LLC, President/Founder, When the AI System Has to Be Right: Healthcare, AV, Policy, Energy. Co-Author of 2030: A Blueprint for Humanity's Exponential Leap

4 年

"Since fatalities occur about every 80 million miles of driving, and 180M miles of freeway driving, 6M miles isn’t enough, it can be argued, to say anything about fatality rates. A?2016 article from Rand?argued that you would need to drive billions of miles to truly prove superior safety. (Rand released?a new study this week?toning that message down somewhat.) "https://www.forbes.com/sites/bradtempleton/2020/10/30/waymo-data-shows-incredible-safety-record--they-should-deploy-today/?utm_source=dlvr.it&utm_medium=linkedin&sh=7f61799d3829 https://www.rand.org/pubs/research_reports/RRA569-1.html

回复
Doug Hohulin

To Save 1 Billion Lives with AI, Exponential Blueprint Consulting LLC, President/Founder, When the AI System Has to Be Right: Healthcare, AV, Policy, Energy. Co-Author of 2030: A Blueprint for Humanity's Exponential Leap

4 年

Nationally, 6.1 million miles of driving by a good driver should result in about 40-60 events, most of which are small dings, 22-27 or which would involve an insurance claim, 12 which would get reported to police and 6 injury crashes. With no at-fault events in 8 lifetimes of human driving,?Waymo’s performance is significantly superior to a human,?even in an easy place like Chandler. https://www.forbes.com/sites/bradtempleton/2020/10/30/waymo-data-shows-incredible-safety-record--they-should-deploy-today/?utm_source=dlvr.it&utm_medium=linkedin&sh=7f61799d3829

回复

要查看或添加评论,请登录

Doug Hohulin的更多文章

社区洞察

其他会员也浏览了