What is a sensible fatality target for AV's - and who sets it?

As AV's get very close to real world bulk testing I keep wondering what the safety goal for them should be - and who should set it. Statements like “anything to stop the daily slaughter on our roads" makes good headlines but no investment or safety engineering sense, (or even a moral sense to the family of somebody killed by faulty AV).

Going back in time when the Royal Aircraft Establishment developed auto landing in the 1960's they had a clear goal - to make auto landing ten times safer than a piloted landing in low visibility and they tested until they achieve their confidence limits on that target. I am sure there are learned articles and confidential company goals on AV safety but I can’t find any in the public media and you would like to think that “informed consent" would apply to AV safety too.

Setting a target is difficult and I would suggest only fatality goals are worthwhile, partially because injuries can vary hugely and also serious injury versus death is probably a function of the base car safety features rather than AV vs. human control.

So what target to set?  You have to ask which country first. The USA has a road fatality rate of 1.16 per 100 million miles versus 0.2 in the UK so what might be an acceptable target in the USA would not reduce UK road deaths. The UK rate represents 816 fatalities in a population of 65 million. Pedestrian fatalities  were 446 some of which could have been due to avoidable driver error. So a round number of 1,000 fatalities potentially avoidable with AV's might be sensible. Total vehicle mileage is 305 billion excluding trucks so one fatality per 300 million miles with human drivers.

So if human drivers are achieving 300 million miles per death what is a good AV target? Obviously they should be better or why spend all that money. The estimates are not going to be precise to some margin or multiple of 300 million is logical. The RAE used X10. Individual AV failures are less catastrophic than planes so maybe X 3 is OK. That gives nice round AV fatality target of one death per one billion miles traveled for the UK.

At that point the obvious problem arises - how can you verify a one in a billion failure and how many failures do you need to experience before you know you have reached your safety target of AV mass usage?

I am nor statistician or safety engineer so I have no idea how you do the testing but a) I think we need a clear target and b)  that target is very, very low as human drivers are , in fatality  terms in a country like the UK actually very safe when driving modern cars.

I'm not sure you do need a clear target. What's the objective? If you must have a target, then it should simply be no worse than for existing vehicles. Your example target, insisting AVs are three times safer, could create a situation where they're banned from use because they'd only prevent half the current fatalities. That would be bonkers.

There are, I think, two questions. 

Why spend billions if they are n beter than human drivers/ Taht investment could have gone into something else to reduce  road deaths or CO2 or whatever.

As the failure rates are very low you have to have target which give you reasonable confidence they are not actually worse. 

I think an order of magnitude reduction in overall fatalities would be enough to offset the outrage over the inevitable few that occur due to AV faults.

I agree that an order of magnitude in a jurisdiction like the US would be a sensible target. If a particular technology were to achieve that, it would no doubt translate to other countries eg the UK - perhaps at some other level - 1/3 X or 1/20 X fatalities.

Of course this is most likely to be self regulated. The manufacturer will need to convince the consumers that their product is safe and road crash statistics will influence the market's choices once AV's become ubiquitous.

I see reduction of fatalities as a no-brainer. Not only will mature AV tech reduce accidents overall, it will also reduce the severity in many cases by "driving to conditions" (and road rules) and avoiding or ameliorating a large number of the more serious accidents that result from drivers taking obvious risks.

The stats is reasonably straightforward. We use a process called a Weibull analysis to look at fatigue tests as they are in progress. This gives early insight into the failure rate of the items under test even though not many have failed. Obviously to get an accurate answer we have to run the full sample, but we don't have time to wait for every test to finish.

Road fatality rates vary considerably between countries with similar cultures and road conditions. 

Numbers are deaths per billion vehicle-km

Austria   5.1

Belgium   7.3

Czech Rep   11.5

Denmark   3.9

France   5.8

Germany   4.2

Netherlands   4.7

Switzerland   3.2

(USA   7.3)

(Australia   5.2)

The most striking differences are between Belgium and its neighbours, and between the Czech Republic and everywhere else.

Source: Wikipedia, from WHO data.

https://en.wikipedia...ated_death_rate

Compared to the UK number posted earlier and the developing countries at the other end of the scale, I don't see anything "striking" on that list. Ten countries, all within the same order of magnitude.

Edited by gruntguru, 06 March 2019 - 22:00.

Road fatality rates vary considerably between countries with similar cultures and road conditions. 

 

Numbers are deaths per billion vehicle-km

Austria   5.1

Belgium   7.3

Czech Rep   11.5

Denmark   3.9

France   5.8

Germany   4.2

Netherlands   4.7

Switzerland   3.2

(USA   7.3)

(Australia   5.2)

 

The most striking differences are between Belgium and its neighbours, and between the Czech Republic and everywhere else.

 

Source: Wikipedia, from WHO data.

https://en.wikipedia...ated_death_rate

 Nevertheless, it is odd that the Czech Rep is more than double the average of the others.    

Compared to the UK number posted earlier and the developing countries at the other end of the scale, I don't see anything "striking" on that list. Ten countries, all within the same order of magnitude.

Really? Death rate in Belgium is 50% higher than Germany or the Netherlands? Assuming that the methodology is the same for all three countries, the rate for Belgium requires a factor to explain the difference. 

If a particular technology were to achieve that, it would no doubt translate to other countries eg the UK - perhaps at some other level - 1/3 X or 1/20 X fatalities.

I disagree. Social or geographical characteristics of countries are a bit different, and road fatality rates are more different -- see above comments from me. Death on the road is more likely in some countries than others which are just like them. Until we can describe the fatality rate differences for conventional cars, we have problems setting rules for AVs.

Self insurance by manufacturers may sort that out. The insurance policy -- part of the up-front cost of the car or a lease agreement -- will cover your old fashioned insurance plus a new liability for whatever the AV might do wrong. The car manufacturer has an incentive to succeed.

Or manufacturers have an incentive to be a bit better than the rest, for a while. 

 Nevertheless, it is odd that the Czech Rep is more than double the average of the others.    

Perhaps we should cross-reference these figures against average beer consumption per country.

The WHO data cited has two columns - per 100K population and per 100K vehicles. The more meaningful rate for AV's is per 100k vehicles as the level of car ownership has little to do with vehicle safety.

Other factors affect the death rates too.

- Quality of emergency care. Better paramedic trainin g has reduced UK death rates and poorer counties often have worse hospital services

- Age of a nations vehicles. Death rates are much lower in newer cars due, mainly to huge passive safety gains.

-Average age of population . For whatever reason younger drivers have more accidents so average age has an effect. 

When Tesla claimed after the "autopilot" crash its car were safer then US average it was being very disingenuous because the average US auto is 12 yrs old and driven by a 30 odd year old. Tesla's are all new with the latest safety standards and driven by  older, lower accident rate drivers who can afford an $80K auto.

The other , huge factor is occupant deaths versus all road deaths including bikers, pedestrians etc. I used 50% of non-occupant deaths in my post as a guess because some are car related but you cant stop  bikers coming of off at corners etc. I don't know if 50% is right but you have to make some estimate.

Oddly the USA ought to have a LOWER total road death rate per 100K vehicles vs most European counties because there are fewer pedestrians and cyclists in with traffic in the USA. so 10.9 in the USA  vs 2.9 in the  UK is statistically significant.

Edited by mariner, 07 March 2019 - 17:51.

Really? Death rate in Belgium is 50% higher than Germany or the Netherlands? Assuming that the methodology is the same for all three countries, the rate for Belgium requires a factor to explain the difference. 

Stella Artois and/or Abbaye des Rocs, accentuated by the numbers of non-Belgians on their roads...

UK about 2.3 ish, fatalities down nearly 30% in the last 10 years apparently.

Edited by Bloggsworth, 07 March 2019 - 19:02.

I disagree. Social or geographical characteristics of countries are a bit different, and road fatality rates are more different -- see above comments from me. Death on the road is more likely in some countries than others which are just like them. Until we can describe the fatality rate differences for conventional cars, we have problems setting rules for AVs.

It doesn't matter what is causing the differences. They include things like:

 - Land area and population density. Big countries with low population density will have more km of B roads with higher speed limits.

 - Simple legislated measures eg in Australia, the road toll was decimated by compulsory seat belt laws. It happened again with random breath testing and 0.05 alcohol limits.

 - Quality of road infrastructure

 - Age of fleet

 - Anything else mentioned by Mariner above.

None of these differences will prevent the benefits of AV technology from reducing the death toll in the areas of driver awareness, compliance with road rules etc - regardless of local factors. The only factor of great significance would be affluence and its effect on uptake.

Edited by gruntguru, 07 March 2019 - 22:57.