Tag Archives: autonomous

Suddenly, Driverless Cars Hit Bumps

Recent tests by The Insurance Institute for Highway Safety on two key ADAS capabilities cast doubt on the efficacy of these technologies and thus on how soon full autonomy is likely to affect auto insurance premium.

Anyone insuring automobiles is paying a lot of attention to the development of ADAS (advanced driver assistance systems) and of fully autonomous vehicles.

Many of the underlying technolgies used in ADAS (e.g. cameras, radar, lidar, AI) will also be used in fully autonomous vehicles. However, the demands that a fully autonomous vehicle places on these technologies are quite different than the demands of an ADAS-equipped vehicle. ADAS-equipped vehicles will pass control to and from human drivers (or send warnings to human drivers) in various circumstances. Fully autonomous vehicles will have no hand-offs and no warnings because there are no human drivers to receive them.

The Insurance Institute for Highway Safety (IIHS) recently ran a series of tests of two key ADAS capabilities: adaptive cruise control (ACC) and active lane keeping. ACC maintains a set speed and a specified distance from a car in front of the car with ACC. Active lane keeping automatically maintains the car within its current lane.

See also: Autonomous Vehicles: Truly Imminent?  

Vehicles with ACC and active lane keeping are at Level 2 on the SAE International scale. This is a widely recognized framework demarcating degrees of autonomy — ranging from Level 0 (no automation) to Level 5 (fully autonomous).

Source: NHTSA https://www.nhtsa.gov/technology-innovation/automated-vehicles-safety

Notice that Level 2 is a long way from Level 5.

The IIHS tested five well-regarded vehicles:

  • A 2017 BMW 5-series with “Driving Assistant Plus”
  • A 2017 Mercedes-Benz E Class with “Drive Pilot”
  • A 2018 Tesla Model 3 and a 2016 Model S with “Autopilot” (software versions 8.1 and 7.1, respectively)
  • A 2018 Volvo S90 with “Pilot Assist.”

The results of these tests were reported in IIHS and HLDI publication, Status Report (Aug. 7, 2018).

The results were not pretty.

  • In one test on a public roadway, the Mercedes was aware of a stationary vehicle in front of it but continued without reducing speed, until the human driver applied the brakes.
  • In a 180-mile test drive, the Tesla Model 3 slowed without an appropriate cause 12 times (including seven instances of tree shadows on the road).
  • In testing active lane keeping on curves; the BMW, the Mercedes and the Volvo were unable to stay in their lane without the driver providing steering assistance.
  • The vehicles’ active lane keeping capability was also tested when they reached the top of hills. At the top, some cars’ technologies essentially lost sight of the lane markings on the road. The BMW failed to stay in its proper lane (without driver intervention) in all 14 tests. The Volvo stayed in the lane in nine of 16 tests. The Tesla Model S swerved right and left as it attempted to locate the appropriate lane. Sometimes it also entered an adjacent lane or drove onto the shoulder.

There is evidence that ADAS technologies do reduce accidents and insured losses—here and here.

See also: Autonomous Vehicles: ‘The Trolley Problem’  

However, the real world test results of Level 2 technology in these five highly regarded models were certainly disappointing. Level 2 autonomy requires the driver to remain engaged and constantly monitor the environment. The key words are “remain engaged.” People, while driving, often do many things other than remaining engaged.

Conclusion

The shared responsibility between less-than-perfect humans and less-than-perfect technologies of Level 2 implies that either the technologies have to become intrinsically better — or they must find ways to compensate for imperfect humans.

As mentioned, you cannot make a straight-line projection of elapsed time from the current state of Level 2 ADAS technology to the arrival of ready-for-prime-time Level 5 fully autonomous technology.

How to Picture the Future of Driverless

Picture this:

The year is 2025. A call comes to the police station—someone has broken into a local home. A drone is deployed to the address and arrives within five minutes. The drone feeds video to the station and to the closest autonomous (driverless) police vehicle. The drone guides the police car to the location. The officer in the car (we’ll assume he’s human, for now!) isn’t actually driving; he’s an occupant, watching the drone’s video feed. He can see the suspect fleeing, and he researches other crimes in the neighborhood along with potential suspects. The drone estimates the perp’s height and weight, and the officer can see his clothing and a possible gun in his belt. The police officer communicates with other officers in the area to coordinate the capture. As the suspect runs, his description and location is fed constantly to all nearby police vehicles, and he is surrounded within 15 minutes of the initial call.

This is far from fiction. The international consulting firm Frost and Sullivan predicts that 180,000 driverless cars will hit the U.S. market in 2020. That’s less than 1% of today’s annual new car market, but that’s just the beginning!

Just about every major car manufacturer (as well as Google, of course) is developing autonomous vehicles, and the competition is getting  more intense as the demand for collision avoidance features grows. Just as drones are spreading (if not yet regulated), driverless cars will become widely accepted. Americans love to drive, but there are too many undeniable advantages to autonomous cars.

The first one is safety. According to the U.S. Insurance Institute for Highway Safety  (IIHS), 94% of all car accidents are caused by human error. Nearly two million crashes could be avoided if human error were eliminated. That’s not to say that driverless vehicles won’t crash, but, as the technology improves, crash rates will drop like a rock. In 2025, if our roads are still packed with commuters, the occupants of many vehicles will be reading, answering emails, video conferencing and browsing the web. In other words, they’ll be working. A recent Morgan Stanley report predicted that driverless cars could add $5.6 trillion (yes, with a ‘T’) to the global economy because of the combination of a steep reduction in accidents and the dramatic increase in productivity. It is estimated that in 2035 autonomous cars will account for 25% of all cars.

Back to the police force. As driverless cars evolve, routine traffic monitoring will drop, high-speed chases will slowly decline (with drone help) and smaller police forces will focus on more serious crime. Cameras will capture everything—both from the ground and the sky. Officers will become highly trained in electronic law enforcement. Efficiency will rule!

Of course, these are just predicted outcomes. This policing panacea isn’t all roses; it will not eliminate the need for community relationships, direct contact with neighborhoods and personal contact in law enforcement. Furthermore, while vehicle collisions will fall, the cost and maintenance of autonomous cars will remain extremely expensive in the near future. Currently, it costs about $150,000 to equip a driverless car. But that cost will drop to $7,000 by 2030 and to $3,000 by 2035.

Nothing’s perfect. Every emerging concept or technology brings unexpected challenges and unintended consequences. But it appears that autonomous automobiles will emerge soon, and it’s likely that some day we’ll say they are “here to stay.”

For today, I guess I’ll have to drive myself home. What a chore.

What I Learned at Google (Part 2)

We didn’t intend to write a series on the symposium that Insurance Thought Leadership hosted at Google last week for C-suite executives of major companies and for regulators, but I want to build on the wonderful post yesterday by Iowa Insurance Commissioner Nick Gerhart, about the insights he picked up there. For me, the symposium underscored a crucial point about the pace of innovation — how it can be faster than we expect at times but can also be slower.

And it’s crucial to get the timing right.

The faster-than-expected part comes from a partner at one of the major Silicon Valley venture capital firms, which we visited as part of the symposium. All these firms track where entrepreneurs are seeing possibilities and where investments are happening, and the partner said that in all of 2014 the firm had been visited by exactly zero people hoping to innovate in insurance. Yet, just in the fourth quarter of 2015, the firm met with 60 companies looking to innovate in insurance.

Even as innovation has surged in fintech, in general, investment in insurtech start-ups has been minimal, about 1% of the total for fintech. But that may now be changing. Start-ups may accelerate the disruption in insurance.

You’ve been warned.

The slower-than-expected (at least for me) part comes from a consensus about driverless cars at the symposium. The group discussions at all five tables reached almost identical conclusions: that fully driverless cars will be feasible technologically in roughly four years but that it will be 10 before they are a major presence on the road.

In Silicon Valley-speak, saying something is 10 years out means it verges on science fiction. After all, 10 years at a pace set by Moore’s Law means that you have some 30 times as much computing power available to you at no increase in cost — if you need that much more power to make something happen, it’s hard to know for sure that it works 10 years ahead of time.

But the concerns of the insurance C-suiters and the regulators were more prosaic. They felt that anyone who might be left behind because of driverless technology would kick up a fuss and that state governments, likely led by the legislatures, could intervene on behalf of constituents to slow the transition.

Perhaps insurance agents would fear the shift of auto insurance from a personal responsibility to a corporate one, shouldered by the manufacturers of the driverless cars or by operators of fleets of the cars — if no person is involved in driving, how can an agent sell personal lines insurance?

Maybe car dealers, already fighting a rear guard action to prevent direct sales by manufacturers to consumers, would fear further loss of their intermediary role — why would a fleet operator need a dealer to purchase of tens of thousands of cars?

Basically, think of anyone who might lose business because of driverless cars and the promised reduction in accidents — parking garages, emergency rooms, whatever — and you can see an obstacle. Not everyone will be explicit about their complaints. It’s hard for an operator of prisons or funeral homes to demand more business. But our discussion groups were sure that opposition would surface in lots of ways and that politicians, always running for reelection, would lend support.

In fact, some technical concerns about driverless cars have surfaced in recent months. It turns out that Google cars have more accidents than human drivers do, albeit only minor accidents thus far and, most importantly, not because of any fault by Google — careless people seem to bump into Google cars a lot at stoplights. Google also acknowledges that the cars would have caused at least some accidents if not for intervention by the highly trained humans sitting in the driver’s seat. So, the technology still has a ways to go.

The pace of technical progress has still been faster than I expected when Chunka Mui and I published Driverless Cars: Trillions Are Up for Grabs nearly three years ago, and we staked out what was then a very aggressive position. The federal government recently stepped on the gas, if you will, by announcing a plan to spend $4 billion on driverless technology over the next decade and to reduce regulatory hurdles for adoption. The rationale — which we have long predicted the government would have to adopt — is that 25,000 lives could have been saved last year on U.S. highways if a mature form of the technology had been in use.

For me, then, the fundamental question from our symposium is: How do you position yourself for a technology that may be wildly important, yet whose timing is uncertain?

Two thoughts:

–A line that carries considerable currency in Silicon Valley is: “Never confuse a clear view with a short distance.” Even if you’re sure that something will happen as part of the transition to autonomous vehicles, keep in mind the issue of timing.

–Then think big, start small and learn fast — a dictum that just happens to come from another book Chunka and I wrote, The New Killer Apps: How Large Companies Can Out-Innovate Start-UpsThat means you get in the game now, with as big a vision as you can conjure up for yourself or your company. Then you start experimenting to see what works and what doesn’t — while spending extremely little money. You make sure you can kill the experiments as soon as you gather the needed information — no pilot projects allowed, at least not in the early days, and certainly no grand plans to go to market. And you keep iterating until both you and the market are ready. Then you start cashing checks.

Actually, one more thought: Consider coming to the Global Insurance Symposium that Nick and the fine folks in Des Moines (my dad’s hometown) are putting on in late April. Nick is as forward-thinking a regulator as I’ve met, and there will be lots of people there who can help you on your journey, whether that involves driverless cars or something else entirely. I’ll be there….

Now Come Autonomous Trucks

In 2012, nearly 40,000 people were killed on U.S. roads, and 90% of those fatalities were caused by driver error.

Imagine an advanced autonomous system that could avoid those deadly motor vehicle accidents. Even a system that works only on the highway — where the technology has already been developed and where trucks spend the majority of their time — can make a significant difference.

A new report has analyzed the impact of driverless cars on the incidence of fatal traffic accidents and concluded that, by removing human emotions and errors from the equation, we could reduce deaths on the road by 90%. That’s almost 300,000 lives saved each decade in the U.S., and a saving of $190 billion each year in healthcare costs associated with accidents. If you expand this to global figures, driverless cars are set to save 10 million lives per decade.

There are now some trucks on the road that begin to fulfill that promise. Daimler Trucks North America’s “Inspiration” freightliner semi-truck this year became the first legally operated autonomous commercial vehicle operating on U.S. highways.

For now, the Inspiration is basically a limited take on the autonomous truck. The driverless system engages when the truck is on the highway and ramps up speed. It then maintains a safe distance from other vehicles and stays in its own lane.

If the autonomous truck encounters a circumstance it can’t handle (e.g., heavy snow or washed-out lane lines) it will alert the human driver that it’s time for him to take over. But what this technology can do is reduce traffic accidents, and that’s why I’m pretty excited about the whole thing.

A human driver has limited situational awareness. Autonomous trucks offer an extra set of eyes that continuously monitor a broad range of sensors (e.g., visible and infrared light and acoustic, including ultrasound), both passive and active, with a nearly 360-degree field of view.

Therefore, driverless vehicles can more quickly determine a safe reaction to potential hazards and initiate reactions faster than a human driver. For example, traffic collisions caused by human driver errors such as tailgatingrubbernecking and other forms of distracted or aggressive driving would be eliminated.

Safer and more efficient driving is the motivating force behind this emerging technology. It’s not about catching 40 winks on the highway or watching an episode of your favorite show. As cool as that might be to imagine, no one is replacing the human as the ultimate decision-maker.

Beginning of the End for Car Insurance?

Volvo’s statement last week that it would accept all liability when its cars are in autonomous mode takes the threat to traditional auto insurance to a whole new level. Google and Mercedes have already made similar promises, so we now have three major companies saying they will treat certain car accidents as product liability issues and will take on risk that has historically been the responsibility of individual drivers and auto insurers.

For good measure, Tesla just offered a software download that will let real drivers in real Model S cars operate autonomously on real roads.

The future is upon us.

Obviously, this is just the start. In Churchillian terms, we aren’t at the beginning of the end for car insurance, and we aren’t even at the end of the beginning; we’re at the beginning of the beginning.

For the immediate future, there will be zero effect on auto insurers. Only a small number of drivers will be operating their cars autonomously and only for a portion of their time on the road. Tesla isn’t even accepting liability at this point, and auto insurers won’t initially even be asked to adjust their rates to reflect the risk that providers of autonomous technology are taking out of the auto policy equation.

A thoughtful column by Craig Beattie argues that two significant steps still have to happen before much risk for car accidents moves to the product liability side of the ledger. First, courts must sort out the many issues that will be raised when the first unlucky person dies in an accident where an autonomous vehicle is at fault. Second, he says, autonomous cars must be in operation long enough that lack of maintenance, rather than product design, becomes the issue that has an autonomous car cause an accident. Courts will then have to sort through who bears the responsibility for that lack of maintenance.

Although I agree with the first point, about settling key issues in court, I’m not so sure the second is a huge deal. I think vanishingly few people will own autonomous cars once we get through the hybrid phase that Volvo, Mercedes and Tesla are taking us into now, where people can switch into and out of driverless mode in what are otherwise traditional cars. Today, cars sit idle more than 95% of the time, so it’s far more efficient to share cars operated as part of a fleet, rather than pay to have what is usually someone’s most expensive asset, or second-most (after a house), just sit there. A study that Chunka Mui and I cited in our book Driverless Cars: Trillions Are Up for Grabs found that a fleet owner could provide cars to people for 90% less than we pay for car transportation now and still make gobs of money. So I believe that fleets, not individuals, will be responsible for maintenance, removing that as an issue that would be in the province of traditional auto insurance.

I also expect the federal government to get involved at some point. If driverless cars can really reduce the number of traffic deaths on U.S. highways (currently roughly 35,000 a year) by tens of thousands and reduce the number injured in accidents (currently about 2.5 million a year) by many hundreds of thousands, then driverless cars create a clear societal good, and their use should be encouraged. Even if the government decided to be revenue-neutral, it could take the money it currently spends through Social Security, Medicare, Medicaid, etc. because of auto accidents and could perhaps cover all the liability for accidents caused by autonomous vehicles — and have a lot left over, besides.

Politics will rear its ugly head when it comes to deciding what government should do and how quickly it can act, but it’s hard to run a campaign in favor of injury and death.

So the issue about traditional auto insurance is much less about if it goes away and much more about when.

“When” is a legitimate question. It takes 15 years or more for the full complement of cars on U.S. roads to be replaced, so you could decide that autonomous-car technology won’t really be mature for a few years, then start a clock and count out 15 years to a time when roads will be fully autonomous. That approach takes many people’s calculations to 2030 and beyond — by which time today’s C-suite members will be safely retired.

But many autonomous technologies, such as forward collision avoidance systems and automated braking, can be installed as a retrofit — Autonomoustuff, advised by our friend Guy Fraker, is a notable supplier. And the dynamics of auto accidents and insurance change long before every car becomes autonomous. Many studies say 20% to 25% penetration is plenty to cause major changes.

While I won’t venture a precise guess about the fate of car insurance, I’ll offer an observation: When Chunka and I wrote about driverless cars 2 1/2 years ago, we staked out what was then an extremely aggressive position about how quickly the transition to autonomous vehicles would happen and about how far the ripples would reach, including for auto insurance — and we may be turning out to have been too cautious.