New efficient electrolysis process for direct hydrogen production from biomass; 16.7% of energy required for water electrolysis
GM exploring Mobileye advanced mapping with OnStar Data; production deployment later this year

Virginia Tech report finds national crash rate for conventional vehicles higher than crash rate of self-driving cars

A new report, “Automated Vehicle Crash Rate Comparison Using Naturalistic Data,” performed by the Virginia Tech Transportation Institute and commissioned by Google, shows that the crash rates for self-driving cars are lower than the national crash rate of conventional cars.

Results show that when data is adjusted for unreported crashes and take into account accident severity, the national crash rate for conventional vehicles is higher than the crash rate of self-driving cars.

Up until now, comparisons based on existing data have been incomplete as requirements in each state for police reported crashes differ, and the majority of severe crashes that go unreported. Estimates of unreported rates of crashes have ranged from as little as 15.4% to as much as 59.7%. The result is that the current national crash rate is essentially a low estimate of the actual crash rate. Meanwhile, self-driving cars are required to report every crash, regardless of severity.

The report examines national crash data and data from naturalistic driving studies that closely monitors the on-road experience of 3,300 vehicles driving more than 34 million vehicle miles, to better estimate existing crash rates, and then compares the results to data from Google’s Self-Driving Car program.

Driving safety on public roads was examined in three ways. The total crash rates for the Self-Driving Car and the national population were compared to

  1. Rates reported to the police. The researchers calculated crash rates from the Google Self-Driving Car project per million miles driven, broken down by severity level. These Self-Driving Car rates were compared to rates developed using national databases which draw upon police-reported crashes and rates estimated from the Second Strategic Highway Research Program (SHRP 2) Naturalistic Driving Study (NDS).

  2. Crash rates for different types of roadways. SHRP 2 NDS data were then used to calculate crash rates for three levels of crash severity on different types of roads, broken down by the speed limit and geographic classification (termed “locality” in the study; e.g., urban road, interstate).

  3. Scenarios that give rise to unreported crashes. SHRP 2 NDS data were again used to describe various scenarios related to crashes with no known police report. This analysis considered whether such factors as driver distraction or impairment were involved, or whether these crashes involved rear-end collisions or road departures.

Crashes within the SHRP 2 NDS dataset were ranked according to severity for the referenced event/incident type(s) based on the magnitude of vehicle dynamics (e.g., high Delta-V or acceleration); the presumed amount of property damage (less than or greater than $1,500, airbag deployment); knowledge of human injuries (often unknown in this dataset); and the level of risk posed to the drivers and other road users.

Google Self-Driving Car crashes were also analyzed using the methods developed for the SHRP 2 NDS in order to determine crash severity levels and fault (using these methods, none of the vehicles operating in autonomous mode were deemed at fault in crashes).

SHRP 2 NDS and self-driving car crash rates per million miles at three levels of severity, Level 1 being the worst. Source: VTTI. Click to enlarge.

Key findings include:

  • Adjusted for unreported crashes and accident severity (accidents that fall within the two highest severity levels), the national crash rate estimates of 4.2 crashes per million miles is higher than the crash rates for the Self-Driving Car operating in autonomous mode (3.2 per million miles).

  • The crash rate of conventional vehicles at all levels of severity is higher than the self-driving car crash rates, according to analysis of the Second Strategic Highway Research Program (SHRP 2) Naturalistic Driving Study.

  • Current data suggest that conventional vehicles may have higher rates of more severe crashes than self-driving cars, but given the small overall number of crashes for the self-driving car at these levels, there is insufficient data to draw this conclusion with strong confidence.

    However, there is statistically-significant data that suggest less severe events may happen at significantly lower rates for self-driving cars than conventional vehicles.

  • When the Virginia Tech Transportation Institute, using methods developed for the Second Strategic Highway Research Program Naturalistic Driving Study, analyzed the Google Self-Driving Car events, none of the vehicles operating in autonomous mode were deemed at fault.

  • As self-driving cars continue to be tested and increase their exposure, the uncertainty in their event rates will decrease. This is particularly appropriate to vehicles intended for lower-speed use where less-severe events are the most likely to be encountered by the newer generation of the Self-Driving Car fleet.


Account Deleted

There is an extremely informative article in Fortune from dec 21, 2015 where E. Musk is interviewed about Tesla's roadmap to fully autonomous cars.

Musk says their autopilot system will be gradually upgraded in the coming years and sometime in 2018 it will become fully autonomous in the sense that it should be able to drive the car at any place a human can drive it and with significantly less accidents than possible by an average human driver. Musk expects that after about one year after they start selling these fully autonomous cars Tesla's datacenter that monitors Tesla's cars will have accumulated enough statistics to overwhelmingly prove that when Tesla's cars are operated in autonomous mode they have less accidents and can drive in more conditions than humans can (like extremely heavy rain and snow where humans may not have visibility but where the radars of the autonomous system still function). Musk will use that evidence to convince authorities that they should start the process of making laws that allow for fully self-driving cars without any requirement for a licensed driver in the car or even a steering wheel or a gas pedal. Musk expect different jurisdictions (states and countries) to spend at least 1 years finishing such legislation and some jurisdictions even over 5 years.

Musk does not say it in the interview but it is obvious in my opinion that Tesla will launch a dedicated driverless taxi service as soon as some jurisdictions allow this to happen.

Consequently Tesla's autonomous game plan look like this. In 2018 Tesla will start selling cars with a fully autonomous autopilot option. In 2019 they will have enough evidence from 10s of thousands of cars to convince some authorities in some countries and states to make fully autonomous vehicles without licensed drivers street legal. In 2020 the first state or country will permit such vehicles and Tesla will start their mass-market autonomous taxi service. After that expect extremely fast growth in such services until they become fully global in 2025 and more or less has annihilated sales of any new gassers and diesels on a global scale in 2030. Expect 50% of all know auto-makers that came too late to the self-driving game to be gone by 2030. Oil will still power a rapidly shrinking global fleet of old gassers and diesels but after 2050 that will also be finished.

Forget about fuel cells as they will not happen and forget about mass-market long-distance BEVs that are fully autonomous and cost as little to own and drive as a self-driving taxi that can be used by many more and be build for endurance. That will never happen either. BEV ownership will be for the wealthy and even most of the wealthy may not want that option because it has very few benefits compared to a taxi service and several disadvantages like taking care of maintenance and cleaning yourself.


One more thing. Musk says that making a fully autonomous car is no longer difficult because all the necessary pieces are in place (read sensors, computers, data centers, software tools). It will only take a few thousands Tesla engineers 24 more months to get the job done.


Picture the scene: You’re in a self-driving car and, after turning a corner, find that you are on course for an unavoidable collision with a group of 10 people in the road with walls on either side. Should the car swerve to the side into the wall, likely seriously injuring or killing you, its sole occupant, and saving the group? Or should it make every attempt to stop, knowing full well it will hit the group of people while keeping you safe?


No system is perfect, even the best fully autonomous car will get into an accident. At that point the question will be 'who is at fault.' The lawyers will want to know.


Moving human (poor) drivers to the back seat will reduce accidents, property damages, insurance rates, injuries, health care cost and fatalities.

Replacing current unshaved badly dressed bus and taxis drivers will be more than welcomed.

The next move may be to replace trucks drivers, pilots, train conductors and many police with automated machines.

It is just a question of time!!!


Self-driving car is a misnomer and more of a PR term for media hype and vapor press releases. At this time they are really automobiles with collision avoidance devices, essentially smart sensors. Safe, truly self-driving cars are years away. and will be available when the guidance systems have the logic to identify all, repeat all, the possible variables that can occur and react safely to them. Also, before they are approved for operation, the responsible authorities must be convinced they are reliability and safe...that will by itself take a long time.

Having said all this, I think the progress in crash avoidance cars is outstanding and hopefully will save many lives. One only needs to watch the vehicle crashes on youtube to see how horrible they can be.


Lad is right. Safety and the cost of accidents will decline quickly as avoidance systems are implemented. We don't need new laws for that to happen, and we don't need perfection. Also, by not insisting on fully autonomous vehicles, the lawyers still have a responsible driver to go after in those few cases where there are accidents.


At this stage of the game, I think these results and conclusions are rather premature.


Please google

"Self driving cars having accidents at twice the norm."

"all crashes involving self driving cars in the state of California must be reported"

Then you can do the estimates.

I will accept the accident rate as between 10 and 40/50 times LESS than human driven car accidents.

There is an assumption required as to the number of unreported incidents not involving self driving cars.

All self driving accidents vs. human drivers only if there is an injury.

In NSW. Aus (no self driving permits) the reporting is mandatory for Vehicles that are immobile I.E. need towing away. Or failing to stop failing to provide details.
At least that minimal knowledge should be enough for basic understanding.

I am suprised how misleading these analysis without the full disclosure are on this topic.
No wonder there are so many contradictory opinions.

Account Deleted

@ ai_vin the way Tesla has solved the situation you depicts is that Tesla's autopilot uses a datacenter where each Tesla on autopilot upload a map of the road it passes with instructions for how to navigate it. If the first Tesla that passes the road turning that corner discovers that road visibility is less than the braking length it will attach a driving instruction to turn that corner at a lower speed so that the next Tesla that drives that road on autopilot will do it at a speed that allows it to stop fully and avoid any accident. This datacenter road mapping and driving instructions is already a reality for the 50k and counting Tesla's with autopilot today. Also any accidents that involves a Tesla with autopilot will be better documented with video, radar and ultrasonic images to establish guilt. I other words, self-driving cars will lead to better justice.

Moreover, Tesla is working to enhance the autopilot system so that it can operate with sensor input from all Tesla cars at a particular location in time. This means that the autopilot will be able to take that corner at higher speeds if another Tesla on that road sees that the road it clear.

The future of the auto industry is driven by driverless technology that in turn will enable cost saving vehicle sharing and pollution fee and highly durable BEV transportation of everything from people to containers. Those who did not get that at this year CES will get it the hard way in a couple of years when they lose their current job.

Scott Vandeweghe

My town relies on motor vehicle violation revenue. Self driving cars don't speed, run stop signs or cut people off. We will see if/when this actually happens.


there is a whole other issue that concerns ethicists related to autonomous vehicles potentially including drone that could be used in war (like) activity.
Next generation military aerial drones could be autonomous.

This applies equally to autonomous ground vehicles.
there are serious concerns about the technology from many and disparate organisations and ethicists already with ground piloted or human operated remote aerial vehicles in both the theatre of war or civilian initiated assassination adventures being possible.

There should be recognition of the potential for that to be also relevant to autonomous vehicles whether public or privately owned.


This result is hardly surprising, given that so-called self-driving cars only work in ideal weather conditions on clear roads. The human-controled accident rate is also near zero in similar conditions.

In other words, the headline should be "new cars driving on perfect roads in perfect weather have fewer accidents." This has nothing to do with autonomous tech, other than the fact that it exposes some current limitations.



Town and cities will have 20+ years to change their revenue sources. Eventually, vehicle owners may have o pay for distance travelled.

Future autonomous drive vehicles could automatically supply all the info required on a daily/monthly basis and/or on an as required basis.

Vehicle owner's credit cards or bank accounts could be automatically charged every month or so. A single (or very few) cloud based collection agency could do the job at very low cost.

The comments to this entry are closed.