Tag Archives: autonomous vehicles

New Thinking on Driverless Vehicles

Driverless vehicles reached a milestone last week when General Motors’ Cruise subsidiary began offering fully autonomous rides (as in, without a safety driver) in San Francisco. While Google’s Waymo has been offering fully autonomous rides in the East Valley of Phoenix for more than a year, operating in San Francisco’s compressed, complex environment takes AV ride-hailing offerings to a new level.

So, having reached this milestone, where do AVs now stand?

Not where many of us expected two or three years ago. And they are heading in a different direction than many predicted.

In particular, it’s now clear that Uber-like ride-sharing services using AVs aren’t going to suddenly pop up everywhere, and, as a result, won’t soon usher in personal ownership of AVs. Instead AVs will be used in environments that can be carefully circumscribed, such as for shuttles on college campuses on a fixed route. And the first fully driverless car you see on the road may be… a truck.

Yes, urban driving and ride-sharing services such as those operated by Waymo, Cruise and Aurora have made extraordinary technological strides, but the difficulties of fitting in with human drivers, cyclists and pedestrians have proved to be much trickier than expected. So, having AVs roaming everywhere, in all kinds of weather, at all hours of the day and night simply isn’t yet practical and won’t be soon, especially because the public doesn’t just expect AVs to be safer than human drivers; the public expects AVs to be flawless, or at least awfully close.

As a result, AVs that focus on passengers will show up for now in “geo-fenced” parts of cities that are fully mapped and well-understood and on set routes on corporate and college campuses, where conditions can be controlled. You may also see small autonomous delivery vehicles like this one, which are being designed to operate at slow speed on sidewalks — college kids need to get late night snacks and beer somehow, right?

Demand For These Autonomous Delivery Robots Is Skyrocketing During This  Pandemic

In addition, Amazon is experimenting with AVs that could deliver its ever-growing array of goods — though the AVs will have to negotiate a host of different environments full of human drivers, cyclists and pedestrians and will have to deal with the complexities of the final delivery. After all, somebody or something has to be there to take an item out of the AV.

Walmart is using AVs to shuttle online orders from a “dark store” to a retail outlet, where consumers can pick them up.

All those gains matter, and amassing experience with enough use cases will, over time, allow for the sort of broad rollout of AVs that many have been predicting.

In the meantime, though, many of the AV ventures are pivoting to trucking because the issues with pedestrians and cyclists go away and the traffic is far simpler to navigate. These ventures will avoid cities when possible and just focus on going hub to hub on the outskirts, according to my longtime colleague Chunka Mui, who has been my go-to on autonomous vehicles since he arranged a ride for us in a Google AV in the early 2010s and who is expert in the field.

He says the CEO of one well-funded company told him he could build an autonomous trucking business just within Texas because it’s 800 miles from El Paso to Texarkana, and interstate all the way. Chunka also notes that Atlanta to Chicago is a 700-mile corridor with massive truck traffic “and only something like seven turns.”

Imagine how much cargo could be transported from and to shipping centers on the outskirts of cities without the need for human drivers and without the limits of a 14-hour day that they must abide by for safety reasons. And imagine how much incentive there is for businesses to capture all that business.

The whole transportation sector would benefit because the reduction in the need for drivers on long-haul trucking routes — even a modest number of them — would do a massive amount to address the current shortage of truck drivers, which the American Trucking Association puts at 80,000 in the U.S. AVs would free the long-haul drivers for other tasks — typically closer to home, where drivers would rather be, anyway.

At the other end of the spectrum, on the shortest of short hauls, autonomous trucks could help resolve one of the issues that currently has the supply chain snarled. Ports are so backed up that drivers may have to sit for hours to pick up a load — and they don’t get paid while they sit. What driver just wants to sit at a port, unpaid, especially while the clock is ticking on the 14-hour maximum they can spend in the cab in a day? Well, an autonomous truck would have no issue sitting at a port. It could inch through the port for however long it took, have a container hoisted onto the chassis by a crane and inch its way out of the port, where a human driver could pick it up. Just being autonomous while inside the port boundary would represent enormous progress.

Now, a fair amount of infrastructure would have to be developed to accommodate this hub-based approach to autonomous trucking. You’d need to build centers on the outskirts of cities. You’d need to adapt truck stops — an autonomous truck isn’t going to refuel itself, and you wouldn’t need nearly as many showers or as much coffee and snacks. And the drivers, as well as the many systems that coordinate their activities, would have to fundamentally switch to a model where drivers shuttle trucks from those hubs outside the cities to and from their final destinations rather than going on the road for days or weeks at a time.

From an insurance standpoint, the trend away from ride-sharing and personal ownership relieves the pressure on auto insurers to rethink their whole model. Although it had seemed that personal ownership of cars might end in the foreseeable future — meaning the end of the need for personal auto insurance and a shift to fleet-based insurance or to a product liability approach for the makers of the hardware and software and to the operators — the old model now seems likely to persist for many years.

At the same time, the move toward autonomous trucking on a piecemeal basis will roil that market, creating threats for some and opportunities for many.



P.S. If you’re interested in more on the trend toward autonomous trucking and on how AVs will likely play out in general, check out a book that Chunka and I published in September along with another longtime colleague, Tim Andrews. We lay out a timeline for the next 30 years for transportation, among other technologies, based on the notion that the best way to predict the future is to invent it. The book is “A Brief History of a Perfect Future: Inventing the World We Can Proudly Leave Our Kids by 2050.”

P.P.S. I go all the way back to the late 1960s with driverless vehicles and, along the way, have seen how long developments can take and how unpredictable they can be — but have also seen real progress.

The connection comes because, when I was a kid growing up in Pittsburgh, my father was the chief spokesman for Westinghouse, which built a prototype of what it hoped would be a driverless mass transit system. What was called Skybus operated on a two-mile stretch of raised roadway in a park near the suburb where we lived, and my dad used to take us kids out there from time to time to ride it. There wasn’t much to it. You walked into what looked like a bus and rode it two miles to the end of the track. The bus stopped, then took you back. But there was no driver or other Skybus representative on board — the Skybus was operated remotely — and this was more than 50 years ago, so the experience was still pretty wild.

Skybus never had a chance as a mass transit system. It was much too expensive to build the dedicated roadway, and there just wasn’t enough saving from not needing a driver in the vehicle. But Skybus did turn out to be a precursor to the sorts of driverless airport trams we’ve all ridden in Dallas and many other cities, including Pittsburgh.

My dad also got a Picturephone when AT&T debuted the service in Pittsburgh in the summer of 1970, so I’m thinking he was a much cooler guy than we kids realized at the time. But the history of phones is another story for another day.

A Changing Vision for Driverless Vehicles

As plans for fully autonomous vehicles continue to get pushed back, the near future is beginning to look like it will revolve around a different acronym: more ADAS, less AV.

Autonomous vehicles, or AVs, will provide many of the technology breakthroughs that allow for advances in ADAS, or advanced driver-assistance systems, which will use a host of new sensors and AI to reduce accidents. But the vision of driverless robotaxis carrying us everywhere and making deliveries looks like it will have to wait a bit, except in carefully circumscribed areas — and maybe even there for a while yet.

The shift to ADAS from full AVs should soften the near-term effects on auto insurers, which have feared a loss of business in a world where individuals aren’t responsible for driving. At the same time, the shift may increase the cost of repairing expensive electronics when accidents occur.

The new focus on ADAS is by no means a statement that the full AV revolution won’t happen. The progress by AVs has been nothing short of astounding since DARPA, a research arm of the Department of Defense, offered a $1 million prize in 2004 in a contest among autonomous vehicles on a 150-mile course in the Mojave Desert. Most of the 15 vehicles chosen to participate were basically golf carts with sensors and computers strapped on to them, and more than half didn’t even make it out of sight of the starting line. The farthest any vehicle went was 7.4 miles. Just 17 years later, we have fleets of sleek-looking vehicles traveling city streets using AI and sensors — albeit still with a safety driver behind the wheel in just about all of them.

Progress will continue, too. A Brookings Institution study found that $80 billion flowed into AV technology investments between 2014 and 2017. That’s just the investments announced publicly and, of course, doesn’t count the prior investments or the money that has flooded into the field since 2017.

The issue hasn’t been that the AV technology doesn’t work — in any given situation, an AV will perform better than the vast majority of human drivers. It’s just that the world around AVs has turned out to be more complex than initial plans allowed for. In particular, we humans do lots of unpredictable things as pedestrians and as drivers — and AVs aren’t allowed to make mistakes.

While we wait for full autonomy, though, plenty of opportunities have opened up to make driving safer, a notion underscored by some recent multibillion-dollar price tags on acquisitions of ADAS companies.

Lidar sensors, governed by always-learning AI, can enhance automatic braking systems — and studies have found that cars are already more than 50% less likely to have a rear-end collision if equipped with such a system. Systems that keep cars centered in lanes will also improve as technology designed for full autonomy is deployed.

Increased communications capabilities designed for AVs will allow for better connections with roads and other infrastructure. When I rented a car last week while on vacation at the Jersey shore, I wasn’t sure what the speed limit was at one point, then realized that it was displayed on my dashboard based on some sort of radio signal from a speed limit sign I’d missed. Cars will also be able to better communicate with each other. If a car slams on its brakes, it will be able to alert the stream of cars behind it so they can instantaneously begin braking, too. Further out, AV technology will even let cars communicate with each other in ways that let them essentially see around corners — even if you can’t see that a car is speeding through a red light and might broadside you, many other cars on the road can, and they’ll be able to alert yours to brake and avoid the danger.

Technology developed for autonomous cars may also find earlier uses in autonomous trucks. Many are looking at having them operate in fully driverless mode on freeways, where vehicle traffic is far more predictable than on city roads and where pedestrians aren’t an issue. Human drivers would be staged at freeway exits, to ferry trucks to and from their final destinations and within cities. Makers of self-driving trucks say they can cut freight costs in half by removing the need for drivers on the freeway portion of long-haul routes.

I remain as optimistic as ever about the outlook for AVs. Since Chunka Mui and I wrote a book on driverless cars in 2013, progress was faster than we expected for a time and now is somewhat slower. As often happens with fundamental innovations like AVs, the development isn’t happening in a straight line. We’re winding up with hybrid forms of the technology in both cars and trucks before we get to the full effects. But we’ll get there.



Don’t Look Now, but Here Come Autonomous Trucks

While the focus for years has been on autonomous cars and on what they’ll do for safety, for auto insurance, for our lifestyles and more, a disruption is taking shape in the nearer term: autonomous trucks.

The fear factor has obscured that vision. While it is odd enough to drive down a street in Phoenix and see a Waymo minivan next to you without a driver, it’s hard to imagine anyone setting loose on a highway an 18-wheeler carrying 50,000 pounds without anyone at the wheel.

But we’re close.

If you can get the image of a hulking semi out of your mind, highway driving makes perfect sense. The issues that have slowed deployment of autonomous cars all relate to the vagaries of us humans. The technical problems related to snow, rain, fog, etc. have all been solved. But is that driver going to shove his way into the intersection even though he doesn’t have the right of way, or will he wait for the AV? What about that pedestrian walking against traffic? That bicyclist who seems confused? Does the ball that rolled into the street mean a little kid is about to follow it from behind the double-parked van? But highway driving takes away a huge number of the human variables — no pedestrians, no cyclists, far less merging with other vehicles.

Autonomous trucks can basically just get on a freeway and go straight until they need to get off. They solve a real problem, too. By law, truck drivers can only drive 11 out of every 24 hours. That means trucks, with valuable cargo, sit 13 out of every 24 hours. It also means that trucking companies are always short of drivers. But autonomous trucks would be able to go 24/7, cutting many trips in half and making trucking much more efficient.

The change would have broad implications, including for insurers that cover truck fleets and their cargo and for those that cover workers — some driver jobs would disappear, while others would morph to handle changes in loading and unloading, refueling and more. For instance, some drivers might become specialists in the “first mile” or “last mile,” taking an autonomous rig through complex city traffic out to the freeway or picking a rig up on the freeway and navigating it to its final destination, much as captains who know a harbor have long done with ocean-going cargo ships.

Change won’t happen overnight. Trucks still have to overcome that first-mile/last-mile problem — a high-profile startup shut down last year after trying to have drivers use virtual reality to take control of trucks whenever necessary. Autonomous trucks will also be more complicated mechanically than cars for the foreseeable future. While autonomous cars will all be fully electric, the batteries necessary to run trucks are so heavy that they cut into mileage too greatly, so autonomous trucks will not only need to have enough battery power to run all the sensors and computers but will still require an internal combustion engine.

Still, workarounds are developing, and autonomous trucks are making great progress. For instance, Locomation offers what it calls autonomous relay convoys, which combine two human drivers and two autonomous vehicles. Each human drives a rig to the freeway, where one then takes charge of driving. The other’s rig switches to autonomous mode and follows the lead rig, while the human in the trailing vehicle rests. When the first driver’s 11 hours are up, the rested driver takes over. Whenever the trucks need to split up to go their individual destinations, the respective drivers simply take control.

These sorts of convoys would barely make a dent in the potential of autonomous vehicles, but they do solve a real problem, and they provide a start for the long adoption curve ahead of us.

While the idea of having 50,000 pounds barreling down the freeway without a human at the wheel may still be intimidating, think of it this way: Truck drivers are up so high up that you rarely notice them, unless they’ve done something aggressive and you’re looking for the driver because he’s ticked you off. And AVs are by nature so cautious that you’d almost never try to stare down their driver. So you may not even notice the transition to automated trucks.



Who Is Liable When a Driverless Car Crashes?

Now that truly autonomous vehicles (AVs) are starting to appear on roads, the insurance industry will be called on to perform its usual role as an enabler of innovation: Insurers will quantify the risks and likely cover much of it.

But how should insurers think about the liability for AVs? Will legislatures specify who is responsible for which problems? Will regulators? Will the courts? What principles will guide the decision makers? Where will liability fall?

Using history as a guide, it’s possible to make reasonable guesses at some of the answers.

An interesting analysis in Fortune argues that the courts will set the rules, applying long-standing principles to try to sort through the issues in the new environment.

The process will thus be messy, and some of the arguments made in court will initially be idiosyncratic. The article notes that, in the 1930s and 1940s, people who were hit by hired taxis sometimes sued the passengers rather than the driver or the driver’s employer. That approach never got traction in the courts and seems silly today, but you can be sure that some similarly odd-sounding theories will be tried in AV cases before being discarded.

The article argues that clear principles will gradually emerge. One is obvious: that the manufacturer will be responsible for a clear error, the software equivalent of having a tire fall off a car. But the two other standards were more subtle:

–A court will ask whether the AV performed better than a competent, average driver. That question may not apply just to the circumstances of the accident and the specific system or component that may have been involved in causing a collision but may also be a general question about the performance of the AV versus a human driver. The U.S. National Highway Traffic Safety Administration made that sort of general assessment of safety when it cleared Tesla’s Autopilot system of responsibility for a fatal crash in 2016. The temptation, of course, will be to compare an AV with a perfect driver — aren’t computers supposed to be free of error? Instead, the NHTSA is taking the position that anything that raises the average competence is a societal good. And a comparison to an average driver would be good news for the manufacturers of AVs and for those that insure them.

–The court will also ask whether an AV performed better than an AV did previously in a similar situation. A key promise of AVs is that they are always learning, and not just from an individual car’s experience but from what has happened to every car in the fleet. So, courts will hold manufacturers responsible for not making the same mistake twice.

The potential revenue for insurers from AVs is enormous. A recent report from Accenture and the Stevens Institute of Technology estimates that, even as AVs slash premium for personal auto coverage, product liability will be one of three new revenue streams that will generate $81 billion in premium between now and 2025. (The other two opportunities are in the new cyber risks that come along with AVs and in the potential liabilities associated with the infrastructure that will support AVs.)

The law will take shape slowly. It always does. There will be surprises along the way. There always are. But the size of the product liability opportunity, plus the beginnings of answers on legal principles, suggests that insurers should start working now to be prepared as the opportunity unfolds.

Stay safe.


P.S. Here are the six articles I’d like to highlight from the past week:

OnStar: Next Step for OEM Partnerships

Insurers hope to create a new way to collect driving data that’s easier for the driver than installing a device or downloading an app.

COVID-19 Is No Black Swan

There were clear warnings about COVID from credible institutions. The real issue is how we are going to deal with “grey rhinos.”

ESG: Doing Well by Doing Good

Insurance is at the forefront of the environmental, social and governance movement, which may usher in a Second Age of Enlightenment.

P&C Claims: 4 Themes for the Future

The extraordinary events of 2020 have accelerated four themes: automating operations; AI for insight; augmenting experts; and new ecosystems.

Advice to Early-Stage Startups on Pricing

Your pricing is a marketing tool that announces how you want potential clients to think of your offering.

How AI Transforms Risk Engineering

“AI could contribute to the global economy by 2030, more than the current output of China and India combined.”

New Push on Autonomous Vehicles

As I got set to write this post on Sunday night, I saw a tweet recommending that I go to bed early and get plenty of rest so I can wake up early and do my part in what may be the least productive week at least in American history, if not the world’s.

Maybe we’ll get lucky, and the contours of the results from the U.S. presidential election on Tuesday will be clear enough, quickly enough, that we’ll be able to focus on normal activities and not be distracted by legal fights and protests, but I’m not counting on it. My younger daughter, in her third year of law school in DC, got an email from the administration over the weekend recommending that she stock up on food, medicines and any other essential supplies as though she were going to be snowed in for a week. That feels about right.

With the world perhaps about to go on hold, I’ll just lob in one, quick thought in this week’s Six Things and hope I get to you before we all disappear into Twitter and cable news for however long it takes to sort out the results of the election. The thought is this: While autonomous vehicles have faded from the discussion since an Uber killed pedestrian Elaine Herzberg in March 2018, AVs are being positioned for a resurgence.

Some of the positioning is official and from the rock-solid actors in the autonomous space. Google’s Waymo unit announced last month that it will begin offering fully driverless service (as in, with no safety driver behind the wheel) to customers of its ride-hailing operation in Phoenix. General Motors said that its Cruise unit will begin operating fully driverless vehicles in San Francisco by the end of the year.

Some of the positioning is more speculative. Tesla followed Google/Waymo and GM/Cruise by announcing that it has begun rolling out to customers what it calls Full Self-Driving as part of a beta test that will lead to a general rollout perhaps by year-end. This would be enormous news, if everything goes as advertised, but Elon Musk has long claimed self-driving capabilities that go beyond what his cars can actually do. With the latest announcement, Musk is claiming that his cars will be able to operate autonomously without access to the sort of carefully calibrated maps that Waymo, Cruise and others prepare before letting their vehicles operate in an area and believe are crucial. Musk is also planning to go live even though he doesn’t seem to have done the kind of extensive on-road testing that Waymo and Cruise emphasize. Many in the industry still doubt that Tesla’s basic technology is up to the task of providing full autonomy; conventional wisdom is that autonomous vehicles require lidar — essentially, a laser-based form of radar — while Musk uses cameras but no lidar.

Some of the positioning is subtle — and this is the part that interests me the most, because I think it may have the biggest impact in the long term. In the wake of the accident that killed Herzberg 2 1/2 years ago, AV companies seemed to fall back and regroup. Now, having done far more rigorous testing, AV companies are starting to resurface with the sort of data that could sell the public on the safety of letting a car drive them around with no one sitting behind the wheel.

Waymo, as usual, led the way, with a report last week on the performance of its driverless cars from the beginning of 2019 through the third quarter of 2020 — and the data is impressive. The vehicles were extremely safe, and — importantly, for building long-term trust — Waymo got specific about its record, using National Highway Traffic Safety Administration standards and going well beyond the press release treatment that usually obscures what actually happened.

Waymo reported 6.1 million miles of automated driving, including 65,000 with no safety driver, and 47 “contact events” — consisting of 18 actual collisions and 29 incidents where a safety driver intervened to prevent a collision and where Waymo then simulated what would have happened without a safety driver. That’s one contact every 130,000 miles, in case you’re scoring at home and want to see how your experience compares with that of the Waymo cars.

Thirty of the 47 “events” resulted in no injuries (or were projected to produce none). None produced “severe” or “life-threatening” injuries (or were projected to in the simulations). While the report wasn’t specific in assigning blame, it said that “virtually all” incidents resulted from an error by a human driver in the other car or by a pedestrian/cyclist. In the eight “most severe or potentially severe” incidents — including a car running a red light at 36 mph and T-boning a Waymo vehicle — Waymo said mildly that “road rule violations” by other drivers contributed.

In the one event where blame likely would have been assigned to Waymo, because it would have rear-ended the car in front of its vehicle, the report suggests that the human driver of the other car was instigating the collision — its driver swerved in front of the Waymo car, then slammed on its brakes, the sort of behavior that is often reported about drivers showing resentment of AVs. In any case, the safety driver saw what was happening and prevented a collision, which simulation showed would have occurred at 1mph.

There’s still a long way to go on AVs, but the announcements by Waymo, Cruise and Telsa show that companies are pressing ahead, and the report by Waymo starts to lay the intellectual groundwork for what I think will be a strong claim for public trust. In any case, after 2 1/2 years of lying low, the AV companies are back to making public arguments for themselves, and the insurance industry will have to react — next week, or the week after, or the week after that, once the craziness from the election settles down.

Stay safe.


P.S. Speaking of public trust, if you haven’t voted already, please do so. We’re better as a people, as a nation, the closer we get to 100% participation by eligible voters.

P.P.S. Here are the six articles I’d like to highlight from the past week:

Property Claims: It’s Time for Innovation

Those that solve for the dynamics of the many opportunities are likely to be the future industry leaders.

Driving Into the Future of Telematics

Connected vehicles, and their shared language of data delivered through an exchange, are the future of telematics.

The Future of Underwriting

A survey finds that nearly everyone expects big changes in underwriting. But what will different look and feel like? And how ready are we?

Asia: Latest Source of Opportunities

Think of giants like Ping An; innovations such as TenCent’s Waterdrop; and ecosystem plays such as Rakuten.

State of Commercial Insurance Market

Every company today is different than it was six months ago. All risk profiles have likely changed.

Best AI Tech for P&C Personal Lines

The value rankings indicate that user interaction technologies fueled by AI are at the top of the list for personal lines insurers.