Who will be liable for a driverless car accident?

Driverless cars will evolve from today’s self parking cars to fully autonomous cars in different steps as explained in a previous post “Driverless cars for the next decades in 4 steps”. This will have many impacts on auto insurers.

step 0: today’s self parking feature and Google cars
The fact that some cars can self park themselves do not have any impact on auto insurance, because though the car is self parking itself, it does it at a speed still controlled by the driver. So the driver is the main responsible for his car liability.

step 1: partially autonomous driverless cars
We’ve been told that Mercedes is going to sell a driverless car in 2013. But the car will be driverless for 10 seconds i.e. drivers will be allowed to have their hand off the steering wheel for a maximum of 10 seconds. As far as I read, the driver will still be liable if anything bad happens during the time he has his hands off the wheel.

For the following years, we can reasonably anticipate that these 10 seconds will become 20 then 30 seconds and even a minute. From here let’s ask a series of questions:

• The biggest one being: “What is the limit from where drivers having an accident, when his car is in automated mode, will be able to attack the car manufacturer?”
• Who will decide what the limit is?
• car manufacturers, government or insurers?
• What if the US go for 35 seconds and europe for 25 seconds?
• How do we prove that the accident happened while the car was in driverless mode?
• Do driverless cars need some airplane like black boxes?

step 2: everyone can operate a driverless car
When everyone will be allowed to operate a driverless car, meaning telling it to drive automatically 100% from point A to point B, driverless car “operators” (I guess they are not “drivers” anymore) will not be liable for any accident anymore.

Who will be liable then for driverless car accident? Maybe the car manufacturer will be liable then. The car manufacturer will then transfer the fee to driverless car users. We can also imagine that driverless cars will now be leased and that they will be liable for any incident. Or maybe, will it be an auto repair shop responsible for car maintenance.

Most probably, will it be a mix of all those stakeholders.

Maybe, it won’t have any significant impact because driverless cars are supposed not to have any accident anymore...

Do driverless cars need V2V communication?

Several car manufacturers work on driverless cars. They explore some different ways of doing it. Volvo is one of the most advanced exploring V2V technology. V2V is an acronym for vehicle-to-vehicle communication. It’s about using some dynamic vehicle to vehicle wireless communications to have cars get extra informations from neighbour cars or from sensors on the road itself.

But Do driverless cars need V2V communication?

To establish communication, you need to be at least 2 and share a common way to communicate. There are plenty of road stakeholders with which V2V communication will not be possible. Even if all vehicles shared the same way of communicating, the wild boar crossing the road in front of your car will not be able to communicate with your driverless car, at least not in a wireless way.

And as a matter of fact, today, most the google car and the soon to be released new Mercedes do not rely on V2V but only or mostly on their sensors.

If V2V is to have a future, a standard has to be established as soon as possible, so all vehicle from different manufacturers can communicate between themselves. Why not making it a “chapter” of AUTOSAR?

But my guess, is that by the time a common V2V protocol is established, most driverless cars will be able to drive completely autonomously, and V2V won’t bring anything to the driverless cars world. Sensors such as cameras, lidars and others will do the job anyway.

My feeling is that V2V for driverless cars is a dead end.

But in a constrained limited environment like a factory or on airport premises, this can be a good solution. V2V would be not between vehicles but rather between vehicles and specially equipped roads with wireless beacons (called V2I Vehicle-to-Infrastructure).

The revenge of faster horses

In 2012, a poll initiated by Bosch, asking about driverless cars acceptance among drivers stirred lot of comments last year: Majority of today’s drivers are not ready yet to be “driven” by automated cars.

Should this stop the effort of developing driverless cars? Probably not, if we remember 2 quotes from some famous innovators, and one from conventional wisdom:

• Steve Jobs -- BusinessWeek, May 25 1998: "A lot of times, people don't know what they want until you show it to them."
• Henry Ford: "If I had asked my customers what they wanted, they'd have asked for a faster horse."
• A banker advising Henry Ford's lawyer not to invest in the Ford Motor Co: "The horse is here to stay, but the automobile is a novelty - a fad"

Maybe drivers not wanting to be driven by automated cars do not accept to just watch the car driving, but I guess most of them, will quickly find much more interesting things to do than driving, such as reading, talking on the phone, watching news... I guess no one likes to drive in traffic jams.

To reflect back on the 3rd quote, actually, “horse did stay”, except that nowadays most of the time riding horse is a leisure.

Maybe in a few decades, driving a car will be a weekend pastime only allowed on some closed loop. No one will be allowed to drive on standard road because it will be deemed much too dangerous.

Maybe, there will be more people riding horse on weekends than people driving cars. It will be then the revenge of the horse...

Driverless cars for the next decades in 4 steps

“Roma die uno non aedificata est” (Rome wasn’t built in 1 day)

Driverless cars will become reality during the following years or decades.

We can classify this evolution in 4 steps:

• step 0: today’s self parking feature and Google cars
• step 1: partially autonomous driverless cars
• step 2: everyone can operate a driverless car
• step 3: shared driverless cars

step 0: today’s self parking feature and Google cars

Technically cars can sometimes self drive themselves in a static environment for a reasonable price: When parking themselves. This is legally allowed worldwide because the driver still controls the speed at which self parking is occurring.

Technically cars can self drive themselves in all dynamic environments for an unreasonable price. This is legally allowed in some US states but a licenced driver needs to be present at the driving seat.

step 1: partially autonomous driverless cars

This step can be divided in several sub steps which are different only in the degree of involvement drivers will be required to have or freedom that they will have. It is the last step where technical issues are involved.

Next year’s Mercedes-Benz: having hands off the wheel for 10 seconds

Technically cars can self drive themselves in some dynamic environment for an “affordable?” price: on highways during stable conditions and in traffic jams. This may be legally allowed worldwide because the driver is allowed to not touch the steering wheel for 10 seconds maximum.

Why did Mercedes-Benz chose that the maximum allowed was 10 seconds and not 5 or 20 seconds? This question is triggering the next sub step.

1 minute (and more) free on highways and traffic jams

“How long end drivers are allowed (technically and legally) not to pay attention to road” will be the most interesting thing to watch for the next 5 to 10 years and especially:

1. environmental conditions allowing it.
2. the price of the technical features needed.
3. reliability.

Driverless cars are autonomous during the entire highway part of a trip

End to end autonomous but car has both manual and self driving mode but a licenced driver is still required to be at the helm.

In the 2 last steps, no technical improvement happens anymore. These steps are about business and environmental changes.

step 2: everyone can operate a driverless car

This step will be reached when it will be not required to have a driving licence to operate a car. All technical issues have been solved.

As in previous step, maybe unlicenced driving will be allowed on some specific, secured, well known trips or when some specific conditions are met (i.e. at specific hours when overall traffic is light).

If driverless cars can be completely autonomous on all or most trips and operated by unlicensed driver, the several features of today’s cars are not needed anymore such as steering wheel, pedals, mirrors... Maybe the driving console can also disappear and the only device you need is a smartphone to enter the trip and poll the time of arrival.

step 3: shared driverless cars

Compared to previous steps, there is no technical improvement any more. Only cab companies are now allowed to own cars and have them maintained.

Shared usage step:

• has little legal implication, except the driver becomes a mere passenger with no responsability
• transforms the way car manufacturers, insurers, auto repairs and cab companies work and interact with each others.
• transforms the way cities are organized

Cars are now as safe as trains. Liability has shifted from the driver to the car manufacturer. Auto insurance clients are no more end users but rather auto manufacturers, auto repair shops or and cab companies.
Parked cars no longer clutter streets and supermarkets parking lots.

You order the car you need for shopping at the local supermarket or for several hours vacation trip with the whole family.

To go shopping:

1. A car that you ordered picks you at home and drives you to the supermarket.
2. At the supermarket, you leave the car and it is then used by someone else or parked away.
3. When you have finished your shopping, the same car or another ones comes picking you at supermarket exit and drives you home.

Everyone becomes James Bond in "Tomorrow never dies" where he drives a car with a remote.

3 driverless cars trends in 2012

To initiate the 1st post of this blog, I reviewed many news coming from the web around driverless or automated cars. I will also develop most of these trends in further blog posts during the following months.

I singled out 3 trends or hot topics of the year 2012:

• automotive manufacturers visions.
• end user acceptance.
• shortly analyzing which actors are leading the driverless car effort.

Automotive manufacturers communicating on driverless cars and putting this at the center of their strategy: Volvo, Continental and Mercedes.
Their visions:

• “no one is killed or injured in a new Volvo by 2020”
• Continental: “will drive us toward a safer, more efficient and more comfortable future”

Safety seems to be the common denominator for automotive manufacturers to develop the driverless car technology. In a later blog post we’ll review all the possible motives for building driverless cars.

A poll initiated by Bosch, asking about driverless cars acceptance among drivers stirred lot of comments last year: Majority of today’s drivers are not ready yet to be driven by automated cars. Should this stop the effort of developing them?

Who is leading the driverless cars effort?

My feeling is that Mercedes-Benz is leading the way towards a driverless future, even more than Google and other players because next year Mercedes-Benz plans to actually sell a new S class Mercedes that allows the driver to take their hands off the wheel for 10 seconds. These 10 seconds make the difference between research and applied science for everyone (I mean, those who are wealthy enough now to buy this car).

Players are german manufacturers (Mercedes, Continental), US with Google. Volvo is also a major player. Seems that China is also doing something (Driverless car completes 114km journey from Beijing to Tianjin.). What about french, japanese and other automotive manufacturers? As of today, they don’t seem to be very active.

Update: Japanese Komatsu has some driverless trucks working in a Rio Tinto's mine in Australia.

Driverless Cars Blog Presentation

I don’t like cars and I usually hate driving, that is why I am passionate by cars and I even work in automotive industry.

I now own a Dacia Sandero that I have to drive from my 300 inhabitants (150 cows and 150 people) mountain hamlet. This is important because, I am constantly relating the driverless car effort to how my daily life could look like when being driven by an automated car.

This blog will features articles about a breakthrough automotive innovation currently ongoing: driverless cars. This blog will deal with driverless cars from multiple angles as driverless cars will not only change the way most human being experience travelling, but will also have tremendous side effects on cities, countryside and business.

Here is a non exhaustive lists of possible way to deal with driveless cars, each linked to a keyword:

• user stories: you as an actual end user interacting with driveless cars [Driverless cars user stories]
• how end user acceptance will evolve until wide adoption [Driverless cars end user acceptance]
• the way business and our environment will change as direct or side effect from driverless cars use [Driverless cars business and environment change]
• technical subjects [Driverless cars technology]
• past, current and future trends  [Driverless cars trends]

About me

I Eric Mariacher, am an average driver, who is longing for driverless car to unload me for the driving task that I don’t especially appreciate. I work as an engineer in embedded software in the automotive area. A few years ago I was involved in major embedded software automotive effort of standardization: AUTOSAR. I also widened my experience by working,as a consultant or employed, in other embedded software fields such as telecommunication, consumer electronics and medical devices.