Speech Transcript: Autonomous Vehicles and the Impact on Fleet

This is a presentation I gave at Utility Fleet Conference 2017 at the International Construction and Utility Equipment Exposition (ICUEE) in Louisville, Ky. on October 3, 20017. Here is the link to the full deck on Slideshare. And below is the transcript...

As a journalist, who has covered the fleet industry for over a decade, I spend a lot of time talking with smart people about the intersection of technology and transportation—specifically, the rise of vehicle automation and how that might impact our world in fleet.

And that’s what we’re going to focus on in today’s talk, as I share with you what I’ve learned from my conversations and research.

Our story begins in 2009 ...

Our story begins in 2009.

It was only 8 years ago but so much has happened since then.

At that time, 

• We were in the throes of the worst economic crisis since the great depression.
• The idea of “calling an Uber” on your smartphone was still about a year away from happening
• And a niche electric car maker, Tesla, had just received a major cash infusion to pull the struggling company from the brink of bankruptcy.

That’s also when search engine giant Google launched its Self-Driving Car project.

If you recall, in 2009, the idea of robot cars still seemed like science fiction--a long way out in the future. 

And any work being done in this space was primarily funded by the Department of Defense. 

So, what makes Google’s foray into this space so remarkable is that this was a private company willing to put significant resources into what the company describes as a “moon shot.” 

And it was an unprecedented level of commitment by the private sector for an unproven, highly expensive technology.

 But today that commitment is starting to pay off with wide-ranging ramifications.

Only 8 years later, Google’s self-driving car project--now branded as Waymo--has logged over 3 million autonomous miles, mostly on complex streets. 

They’ve also spurred on a lot of healthy competition, with traditional automakers and Silicon Valley companies--including Tesla--racing to launch their own self-driving systems by the beginning of the next decade. 

And our world--and your career in fleet--will never be the same. 

Agenda

So, here’s the plan for our time this afternoon.

First …Besides the fact that this is really cool technology, why should you care? I’ll give you three reasons.

We’ll then define the key terms you need to know to serve as a framework--and give us a common vocabulary--for our discussion.

I’ll outline the Top 6 Trends that I’m seeing in vehicle automation—that make this idea of a self-driving world seem inevitable.

But, then, we’ll also talk about the 3 Biggest Hurdles that could put the brakes on the commercialization and adoption of this technology.

We’ll ask ourselves, based on what we’ve learned today, what could be the impact on fleet? I have some initial ideas to get the conversation started.

We’ll tie this all together with some Bottom Line remarks.

And then we’ll move into a Q&A discussion time. 

Why Care?

Reason #1: Safety

The number 1 reason you should care … Safety.

In his speech at the Automobili-D Conference in Detroit this past January, John Krafcik, the CEO at Waymo cited this compelling statistic: “Each year, more than 1.2 million people die on the roads around the world.”

He then put that number in context: “That’s equivalent to a 737 [airliner] falling from the sky every hour of every day all year long.” 

Krafcik’s point is clear. 

Society would never tolerate having a major airline crash every day--let alone 24 crashes--so, how can we accept the same number of people dying in automotive crashes? 

And, if we know that 94-percent of crashes can be tied to human error. We’ve identified the problem, right? It’s us.

So, if self-driving systems could take the human driver out of the equation--and thus, prevent the vast majority of fatalities on the road--wouldn’t it be a moral imperative for society to adopt that technology?

That’s the argument that Krafcik, along with several Silicon Valley entrepreneurs and most automotive executives have been making in recent months as they present a vision of a “crash-less” society made possible by fully autonomous vehicles. 

Are they right? Will self-driving vehicles usher in a crash-less society?

No one really knows for sure. But what we do know is that “safety” is a powerful argument to spur regulatory action to create rules that govern autonomous vehicle operation on public roads.

Reason #2: Productivity

The second reason why you should care...is productivity. 

The average commute (to-and-from work) for U.S. workers is over 50 minutes.

That equates to about five 40-hour work weeks over the span of a year. 

And you can imagine that most of your equipment operators and technicians spend even more time behind the wheel driving to multiple job sites each day. 

Self-driving vehicles could change that equation by converting all those wasted driving hours into “office time,” with huge implications for worker productivity.

The automakers have already started thinking about this in their future designs, as you can see from reporting from just a couple weeks ago, with Ford’s patent application for a retractable table for autonomous vehicles. The company has reportedly filed patents on retractable pedals and steering wheels, as well. 

But here, we’re only talking about commute time--to and from work. What about impact on productivity on the job? 

This question got me curious, and I had an opportunity not too long ago to explore this idea further in a project I was putting together for the field service industry, which has many similarities to the utility market.

I thought through a scenario and presented it to industry analysts at Navigant Research and Juniper Research who do a lot of work in the autonomous vehicle space to get them to give me a “sanity check.” 

Here’s what I presented to them:

Imagine this scenario: Instead of owning service trucks, you would have on-demand access to a publically or privately owned pool of self-driving pickup or flatbed trucks, where you can schedule those vehicles to arrive at the facility first thing in the morning. Service technicians would set-up and organize their tools in a removable fiberglass capsule or pod, which would be lifted (with autonomous forklifts) onto and off the bed of the self-driving truck. This way, technicians can easily switch from one vehicle to another without tying up their time with reloading tools, equipment and product each time they change to a different truck.

Then imagine that a technician is at a customer site but realizes he doesn’t have a component in his inventory needed to finish the job. Instead of having to leave the customer and schedule another trip back, he can call the warehouse and have the part loaded onto an autonomous truck or “delivery pod” to be delivered directly to him on site.

And instead of completing paperwork while parked outside the customer’s location, the service tech can use the commute time between jobs for administrative tasks. This would shorten the time between dispatches and enable technicians to get more jobs done per day.

So, is this plausible or am I just crazy?

Said David Alexander at Navigant: “Yes, your vision of the future service organization is one possible scenario. It could mean field service companies may no longer need to own or manage a fleet. The independent vehicle fleet would have a selection of vehicles that would drop off technicians and their tool pods at the site and deliver parts when needed.”

Sam Barker with Juniper agreed. “The scenario you suggest is likely to occur. The introduction of autonomous vehicles means we will see a completely different vehicle ownership model with users tending to rent out the vehicles rather than owning one themselves. Users will also see productivity increase as commute times can now be used for other tasks.” 

But Barker did have one caveat. He said that the job of unmanned parts delivery might not be performed by a self-driving ground vehicle, said Barker. “It is more than probable that [an aerial] drone will be used to pick up small items rather than an autonomous vehicle.”

So, when would this type of scenario become a reality?

Reason #3: It's not a matter of if but when ...

The answer to this question brings us to Reason #3 why you should care … It’s not a matter of if but when …

As Dave Alexander predicts: “Fleets are going to take their time to learn about the practicality of self-driving vehicles just as they do with any new technology. The use of the vehicles will be very closely monitored and limited to a small number of selected fleet customers in specific sectors -- such as a taxi service. I would say that 2030 is when autonomous vehicles really become mainstream [in fleets].”

As Juniper Research predicts: 20 million fully autonomous vehicles will be on the road globally by 2025, with consumer adoption set to take off in 2021.

And 2021 is significant because that’s when many of the OEMs expect to be launching their autonomous offerings. 

And then there’s Tesla which is pushing the pace.

At this year’s TED Conference, Chris Anderson who heads up the Conference asked Tesla CEO Elon Musk: “How long until you can get in your car, fall asleep, and wake up at your destination?

Musk: 2 years. …

That’s 2019. 

But that’s when Musk expects it to be available for customers.

In terms of that technology being ready for testing? Make that this year--2017.

At the same TED Conference, Musk said that Tesla is on track for completing a fully autonomous vehicle: “November or December of this year, we should be able to go from a parking lot in California to a parking lot in New York, no controls touched at any point during the entire journey.”

The bottom line here is that this technology is coming--sooner or later.

Key Terms

So, if it’s not a matter of IF but WHEN, then it makes sense that we begin to develop a framework for how to think and talk about this technology. And that’s what we’re going to begin to do by defining the key terms you need to know. 

A metaphor that I’ve developed to help me understand this technology is to compare it to how we as humans see, think and react--with our eyes, brain, and nervous system.

The Eyes: 5 Sensor Technologies

Just as we humans must be able to see to drive, so do self-driving cars … and they see through sensor technologies. Here are the five most prominent being discussed today.

Sensor Technology #1: Lidar

This is the siren-like device that you’ve seen spinning atop the roof of the Google car, with dozens of lasers, generating 360-degree maps of the vehicle’s immediate surroundings.

To give you a frame of reference, the Lidar being used in Ford’s autonomous test vehicles operates at a rate of 2.8 million laser pulses per second, being able to pinpoint precisely where the vehicle is positioned on the 3D map, so that the vehicle can safely navigate its surroundings, even on winding roads in the dark of night, without headlights. 

Sensor Technology #2: Cameras

The advantage to cameras is that they can use optical recognition to not only spot obstacles, but to also interpret what they are in a way that helps the car determine the best course of action. Is it a car? Or a pedestrian? Or a cyclist that appears to be veering into traffic? 

Cameras are also used to detect and interpret lane markings and traffic signs. 

Sensor Technology #3: Radar

Radar emits medium-range radio waves that determine speed and distance of the vehicle ahead--commonly used in collision avoidance systems.

Radar’s limitation is that it cannot “see” lane markings and road signs or identify what type of object is being detected, whether a car, pedestrian, animal, or debris.

Sensor Technology #4: Sonar

 Sonar sensors produce ultrasonic waves to spot obstacles in short-range applications, such as:

• Blind spot detection
• Park assist
• Fully automated parking systems. 

As with radar, sonar cannot interpret what it sees, just that some sort of object has entered its field of “vision.”

Sensor Technology #5: V2X

V2X refers to vehicle-to-(vehicle/ pedestrian/ infrastructure/ everything). 

While other sensors (such as lidar and cameras) require clear line-of-sight, V2X systems, like those being developed by Denso Global, enable cars to “see” other vehicles through wireless communications that your own eyes won’t be able to detect.

It uses short-range wireless communications so that cars, pedestrians (carrying smartphones), traffic lights (infrastructure) and eventually anything with wireless connectivity can all communicate with each other in real-time to ensure smooth and safe traffic flow. 

The Brain: Machine Learning

It’s one thing for an automated vehicle to “see” its surroundings, but how does the machine interpret that information and make decisions that keep you safe? With constantly changing road and traffic conditions, how does the car keep up and adapt? That’s where machine learning comes into play.

So, what exactly is machine learning?

It’s a type of artificial intelligence that can learn and improve decision-making on its own, without having to be programmed to perform specific tasks.

This is especially important because it’s impossible for a human programmer to think through all the infinite possibilities of driving scenarios that we as humans handle intuitively--and then try to program all those scenarios into the system.

But with machine learning, you could develop the system to learn and adapt on its own as it encounters different scenarios—either as a single vehicle or as part of a multi-unit fleet that shares the “learnings” across the entire network of vehicles, making them all smarter, simultaneously.

And the more machines you have on a fleet network, the greater and better the data those machines will generate and learn from. And that’s why autonomous miles driven matter—to train the machines to get smarter and smarter.

The Nervous System: ADAS

The “nervous system” refers to driver assist systems that are designed to enable the car--taking in the data from the “eyes” and processing that data through the “brain”--to then react to dangers faster than your own reflexes.

Driver assistance means that the driver still involved.

(Pointing to the bullets on the slide.) These are examples of ADAS technology that you’re likely familiar with.

The important takeaway here is that today’s automated driver assist systems are the building blocks for tomorrow’s fully autonomous cars. 

What Exactly Does "Autonomous" Mean?

The Federal government adopted the SAE standard last year and it appears to be a global standard as well.

Two key responsibilities to be thinking about when distinguishing the levels:

• Responsibility for the driving task
• Responsibility for monitoring driving environment

The DOT draws the distinction between Levels 0-2 and 3-5 based on whether the human operator or automated system is primarily responsible for monitoring driving environment.

Level 0—human driver does everything. 

Level 1—ADAS can sometimes assist the human driver with a driving task—one ADAS--either steering or acceleration.

Level 2— Automated system can conduct some parts of the driving task, but the driver must be fully engaged with monitoring the environment and performing the rest of the driving tasks. L2 is a combination of ADAS--both steering and acceleration. —Tesla Autopilot

Level 3—Automated system can BOTH perform some parts of the driving task and monitor the driving environment in some instances—but the human driver must still be ready to take back control when the system requests.

Level 4—Automated system can both conduct the driving task and monitor the driving environment—without the need for the human to take back control. BUT it can only operate in certain environments and conditions.

Level 5— Fully autonomous; no restrictions. Automated system can perform all driving tasks, under all conditions that a human could perform them. 

Top 6 Trends

So let’s look at the trends that appear to drive us toward an inevitable self-driving future.

You’ve seen ADAS systems having been deployed in cars the past few years. But now we’re starting to see them being introduced from the factory in the commercial truck market.

Here are a couple examples of what’s coming in 2018 models:

2018 Freightliner Class-8 Cascadia with Detroit Powertrain

• Include adaptive cruise control and collision mitigation with automatic braking.
• They have an Intelligent Powertrain Management system that operates like a “predictive” cruise control, using GPS connectivity that enables the truck to anticipate upcoming road terrain and automatically adjust transmission shifting, engine acceleration, and braking in a way that maximizes fuel economy, as the vehicle approaches each hill, climbs it, and coasts on the other side. 

2018 Ford F-150 

• Features an available Pre-Collision Assist with Pedestrian Warning system and an advanced adaptive cruise control with “stop-and-go functionality” that uses radars and cameras to maintain a set distance behind a vehicle–and even follow that vehicle down to a complete stop. 

And these are just a couple examples. GM, RAM and others have begun expanding ADAS offerings into their trucks as well.

So what does expanded factory ADAS availability have to do with the trajectory of autonomous vehicle development? 

• ADAS systems serve as building block technologies for AVs.
• And the more the OEMs get involved--this creates economies of scale
• Which helps make the key components cheaper and better -- to eventually make AVs financially viable.

The idea here is thinking of the vehicle as the "Ultimate Mobile Device."

And connectivity enables technologies for the vehicle to send and receive data in a way that is shaping the future of autonomous vehicles in these five areas.

1. Over-the-Air Updates

That’s not only to update software—but it can be used to update hardware, as well… 

For example … After Hurricane Irma, Tesla used over-the-air updates to give customers an extra 30-40 miles of range, but made it temporary. The batteries would lose their extra juice by the weekend.

The battery boost applies to Model S and Model X vehicles that were built and sold with 75 kWh battery packs — but were software-limited to 60 kWh or 70 kWh capacities. Customers who bought the cars got them at a lower price, with an offer of paying to unlock the remaining battery power later.

That’s amazing technology.

2. V2X 

With V2X--we’ve defined this earlier. It’s essentially the 6th sense for an autonomous vehicle.

And connectivity is the key to making that happen.

3. Real-Time Mapping

No matter how detailed and current a map might be, actual road conditions can change in an instant. 

That’s what makes the real-time mapping system so compelling for autonomous vehicles. 

As a car travels and encounters a new construction zone, that data is transmitted wirelessly to the cloud and then made available to other vehicles that will be traveling in the same vicinity, automatically updating their maps. 

In other words, the vehicle ‘learns’ as it goes, uploads that data to the cloud and then downloads the next few miles, as needed.

4. Platooning

Connectivity also enables truck platooning, which is emerging as an intermediate commercial use-case for semi-autonomous technologies. 

Just last week at the North American Commercial Vehicle Show in Atlanta, Daimler’s North American truck unit said it received permission from the Oregon Department of Transportation to test its platooning technology on public roads after successful trials in its proving ground in Madras, Oregon.

And in April last year, three autonomous Mercedes-Benz semi trucks completed a cross-border convoy drive from Stuttgart in Germany to Rotterdam in the Netherlands -- about 400 miles -- as a “connected” platoon. 

The trucks were equipped with Daimler’s Highway Pilot Connect, which uses electronic vehicle-to-vehicle (V2V) networking between the trucks, allowing electronic docking -- or platooning -- by vehicles on long-distance highways.

According to Daimler, the connected vehicles in a platoon require a distance of only 15 meters (49 feet) instead of the typical 50 meters (164 feet) between them, which significantly reduces aerodynamic drag, achieving fuel savings up to ten percent.

But here’s the kicker: Remember when we talked about the “nervous system” aspect of ADAS technologies?

According to Daimler, while a human behind the wheel has a reaction time of 1.4 seconds, Highway Pilot Connect transmits braking signals to the vehicles in less than 0.1 seconds. 

And it’s connectivity that makes this possible.

5. Remote Control

Imagine … You’re driving a truck but not sitting in the cab. Instead, you’re at an office or warehouse near your home, several hundred miles from that vehicle, controlling it remotely from a computer, much like how a military drone pilot operates unmanned aerial vehicles from the other side of the globe.

And your job is to “drive” that truck via remote control from the shipper to the highway, at which point the vehicle transitions into full self-driving mode. Then, when the truck is ready to exit the highway, you retake control and guide the vehicle from the off ramp to the delivery destination.

This sounds far-fetched, doesn’t it?

Well, a Silicon Valley startup, Starsky Robotics, has developed a remote control autonomous trucking system to do just this. Earlier this year, a Starsky-equipped Freightliner Cascadia hauled a 5,000-pound load from Orlando to Ft. Lauderdale, traveling 120 miles autonomously on the highway and another 20 miles via remote control on local roads. 

Starsky’s aftermarket retrofit kit includes robotics controls that physically push the pedals, turn the steering wheel, and change gear. And the remote drivers are able to take control of the trucks at any time. 

The company says that each “driver” will be able to monitor and occasionally control between 10 and 30 trucks at a time. 

As you look at all these developments, you can see that vehicle connectivity is a HUGE break-through technology when it comes to making eventual fully autonomous vehicles possible. 

1. Shift in attitudes (and behavior): ownership vs. on-demand.

How many of you have ever taken an Uber or Lyft?

Their vision is on-demand transportation.

And they seem to be making an impact on consumer attitudes and behavior toward vehicle ownership.

According to a Reuters/ Ipsos poll in May, nearly 25% of American adults sold or traded in a vehicle in the past 12 months. 

While most of those people got another car, 9% of that group turned to ride-sharing services like Uber and Lyft as their main form of transportation.

Now, one the big reasons that drove Silicon Valley investors to pour hundreds of millions of dollars into these transportation startups is what they call the inefficiency of car ownership.

Their argument goes like this ...

If the car is driven 2 hours per day, that mean it sits for 22 hours—91% of time.

And with ownership you still have to pay all the on-going costs associated with it.

That’s a highly inefficient economic model. 

The Uber model allows you to pay ONLY for the transportation that you actually need and use.

And the ultimate vision for these ride sharing companies is that they can provide that transportation more efficiently—and at lower cost—through driverless vehicles.

2. New OEM business models--Sales vs. Services

OEMs are starting to see the early impact of this shift in attitudes toward vehicle ownership. And they’re making significant investments in the ride-sharing game—to hedge their bets on a future where they might be forced to transition from a per-unit sales model to an on-demand services model. 

Here are a few examples:

GM has made a $500 million investment in Lyft. And the company has also developed its own proprietary ridesharing platform branded as “Maven.” 

• Daimler Car2go business offers Smart and Mercedes-Benz brand models for short-term, one-way rides. 
• BMW’s ReachNow app offers a similar service for on-demand transportation. 
• Toyota owns a small stake in Uber and is partnering with the company. 
• And just last week Ford announced a partnership with Lyft.

3. Early use cases for the networked autonomous fleets

So what does this potential shift in vehicle ownership attitudes and OEM business models have to do with autonomous vehicles?

Take this quote from GM’s Maven’s chief growth officer in a recent Automotive News article: “You can see a world where you have a Maven platform and plug an autonomous vehicle into it. You can see an autonomous future.”

All this is to say that this emerging trend toward on-demand transportation is providing an early use-case for the commercialization of fully autonomous vehicles. The idea here is that you can build the network with human drivers and then replace the humans by “plugging-in” the autonomous vehicles onto that network--in a way that makes good business sense. 

The big news here is that Waymo brought design and production of all self-driving hardware and software in-house with their Waymo-equipped Chrysler Pacifica Hybrid vans—and this created big break-throughs in cost reductions. 

Waymo CEO John Krafcik in a speech at the Detroit Auto Show in January said that by designing their own lidar, Waymo has been able to improve performance, while taking 90-percent out of the cost--from about $75,000 a few years ago to around $7,500 today. 

So what?

Beyond reducing the cost of sensors right now, if Waymo can spread production of their self-driving systems across multiple automakers and models, this would further drive down the cost, making autonomy more affordable for mass-market growth.

Major automakers--GM, Ford, Daimler, BMW, Volkswagen--are expanding their presence in Silicon Valley and forming relationships with emerging technology startups that could help accelerate development of autonomous vehicle technologies. 

These developments (pointing to slide) are just a few of numerous recent examples.

And there’s something new going on every day. 

My recommendation if you’re interested in keeping up with latest developments in this space is to sign up for Google Alerts under “Autonomous Vehicles.” You will find fresh news EVERY DAY on this topic.

The point here is that as OEMs and venture-backed startups put more and more money into this space, this builds momentum that self-driving vehicles aren’t a novelty or fringe technology but something that’s being taken seriously as a future product offering.

And this will help ensure that there will be more testing and more autonomous miles driven, making these vehicles smarter to the point that they’re ready for commercialization. 

This is important because the technology for self-driving vehicles, for the most part, is already here. What’s needed moving forward is a regulatory framework that makes these vehicles safe and legal.

There are two important developments to note here:

First, just last month, the House of Representatives passed H.R. 3388--the Safely Ensuring Lives Future Development and Research In Vehicle Evolution (SELF Drive) Act. 

Leave it to Congress for being the branding wizards of our time!

Three key things to note with this bill, which was approved on a bipartisan vote:

1. Establishes a national framework for the use of self-driving vehicles
2. Defines the roles of the federal and state governments for self-driving cars, including a requirement for the Department of Transportation (DOT) to develop sweeping regulations for AVs. 
3. Congressional action is now in the Senate’s hands. 

And NHTSA, starting last year--with this document--has been more actively involved with providing guidance for the industry on potential Federal policy on AVs.

And the Trump administration a couple weeks ago released its “2.0” version as you see on the slide.

Without getting too much in the weeds on this, here are the key points to take away as to why this matters for the development of autonomous vehicles. 

1. Congress rarely agrees on anything anymore. The bipartisan support shows cooperation, commitment, and momentum that something will likely get done here.

2. The patchwork of state laws would likely slow down development. So having a national framework would create consistency.

3. And that consistency, driven by the guidelines for development, will allow OEMs to build autonomous vehicles for all 50 states--making it more economical to manufacture the vehicles.

The Biggest Hurdles

Now, we’ve spent the past 6 trends talking about how we’re marching toward this inevitability of a self-driving world. But there are 3 big hurdles, if not addressed properly, could put the brakes on this.

Here’s the deal: Although autonomous vehicles offer the promise of significantly greater safety than their human-driven counterparts, U.S. drivers don’t believe it--at least not from an emotional and practical standpoint. 

That’s based on the findings in a report from AAA earlier this year, where three‐quarters of U.S. drivers said they would be afraid to ride in a self‐driving vehicle. 

And the majority (54 percent) of those drivers said they would feel less safe sharing the road with fully autonomous vehicles while they drive a regular vehicle.

You might think, “O.k., that makes sense when you factor in older generations that may be more apprehensive about new technology, but what about millennials? Certainly, younger people would be much more open to riding self-driving vehicles.”

Yet, according to the AAA study, 73 percent of millennials also indicated that they were likely to be afraid to ride in a self-driving car, compared to 75 percent for Generation X and 85 percent for baby boomers--not that big of difference. 

Now, the industry has taken notice of this issue and are working to counteract this fear. 

Companies like Waymo, Uber, and Boston-based nuTonomy have recently launched programs that offer self-driving rides to select passengers in limited locations around the world. The idea is to get people used to riding in these vehicles and share their experiences with family, friends and colleagues, with the hopes of not only reducing fear but also increasing market demand for self-driving rides.

I began asking the question, what exactly are people afraid of when it comes to autonomous vehicles?

The data in the AAA study doesn’t dig that deep. 

But I know, personally what I would be afraid of.

I can only imagine traveling 70 mph in a closely connected convoy of other autonomous vehicles and then I see this...

The "hourglass of death"! 

I don’t know about you. But for me, the industry needs to have sufficient redundancies in place that would keep this type of “crash” from happening.

But they also need to figure out how to do that in a cost-effective way. And that can be challenging.

A self-driving vehicle is approaching a traffic situation where there will be an unavoidable crash. The car must decide between killing 10 pedestrians or its own passenger. 

What would you say would be the right moral choice?

According to a recent MIT study, 76-percent of participants said that it would be “more moral” for the autonomous vehicle to sacrifice one passenger than kill 10 pedestrians.

This is based on the moral philosophy of utilitarianism where a morally good action is one that helps the greatest number of people--in this case, allowing the vehicle to sacrifice the one passenger to save 10 pedestrians.

But what if you’re the passenger of the self-driving car? 

Now, that’s a different story. 

According to the study, you’re more likely to prefer a vehicle that will protect your life, not sacrifice it. 

“It appears that people praise utilitarian, self-sacrificing [autonomous vehicles] and welcome them on the road, without actually wanting to buy one for themselves,” said the report.

This is a prime example of what the researchers call a “social dilemma,” where people may have a strong consensus on what’s best for society as a whole, but will still prefer to act in their own self-interest. And this double standard could have huge implications in terms of impeding the development of regulations that will make autonomous vehicle commercially available. 

To encourage more public discussion on this issue on a global scale, one of the study’s authors, launched Moral Machine, an online platform that invites the public to get involved with building a crowd-sourced picture of human opinion on how machines should make decisions when faced with moral dilemmas and discussing potential scenarios of moral consequence.

And this brings us to our third big hurdle.

At this point, you’re thinking, "Wait a minute, politics? You just said that political momentum was a positive trend."

And for the most part it is, but there are a couple catches.

And this is all going on right now in Congress. 

• House Bill SELF-DRIVE Act excludes commercial trucks above 10,000 lbs. GVWR
• Senate considering adding commercial trucks—w/ Navistar and ATA advocating for inclusion because the trucking industry and OEMs want some form of guidelines before they make huge investments in these systems.
• But the Teamsters are lobbying against commercial truck inclusion—impact on jobs

And that’s just the stuff that’s going on right now. 

You also have long-term political challenges.

Take for example the idea of trying to achieve legislative agreement on ethical frameworks:

Think about how thorny this issue can be. It hits at the heart of religion, personal values, and moral decision-making. 

It’s hard enough as humans to make split-second moral decisions in crisis. But at least we have the power at that moment to choose with our conscience. 

But would we, as a society, be o.k. with the idea of being spectators inside machines that make life-and-death decisions on our behalf, without our consent?

Engineers are achieving quantum breakthroughs in artificial intelligence that will function as the “brain” for tomorrow’s fully autonomous vehicle. But will they figure out how to give a machine a soul? 

That’s the question--and that’s a big dilemma for Congress and the industry to sort out.

What's the Impact on Fleet?

Assuming that we as a society work through these hurdles and we continue our march toward a self-driving world, what will be the impact on fleet?

I came up with a few ideas…and I would love to get your feedback and also hear what you think.

Driver training—b/w levels of automation

What’s going to happen when you’re managing a mix of different-level autonomous assets?

Vehicle specification and Driver policies

Now that more and more OEMs are offering semi-autonomous systems as factory options, does this mean that you should automatically spec those technologies in the name of safety? Or, is the incremental cost to include those options still too steep for the budget? 

And … what are your company’s policies with operating vehicles with self-driving capabilities? If a driver gets annoyed with the beeps or vibration alerts on the truck’s collision mitigation system and decides to disable it, how is that issue addressed? If a crash occurs after the system was disabled, what does that mean for your company’s risk exposure?

Risk management

And that goes into risk management …

When a robotic vehicle is involved in a collision, who (or what) is responsible? 

There’s a growing consensus around the idea that the automaker would assume liability. But what would happen if a sensor on your self-driving truck failed to detect a child darting behind the vehicle as it shifted in reverse, fatally striking that child? 

Sure, you might have grounds to blame the OEM for the sensor malfunction, but it’s your utility’s logo on that truck. Now, your organization is dealing with a PR firestorm for an incident (and a truck) your crews had no control over.

Fleet vehicle upfits and equipment

Or, what about the liability with upfitted trucks? 

It’s one thing when self-driving sensors are installed by a single car or truck manufacturer. But what about when a third-party upfitter mounts a body on a chassis? Would the upfitter be responsible to install the cameras, sonar, radar and other sensors on the truck body and integrate them with the sensors on the chassis? If so, how does the industry ensure safety and quality control of the self-driving systems for both the chassis and body?

And who, ultimately, would be liable for an incident caused by a sensor malfunction on an upfitted truck? Would it be the chassis OEM, the upfitter, or the sensor manufacturers?  

Or, what if there’s a computer glitch? When a vehicle is driven by software, who would be allowed to work on it? And if persistent system glitches occur, causing a collision, who’s the responsible party -- the vehicle OEM, the repair shop, the utility itself?

Vehicle lifecycles and ownership model

On owned vehicles that theoretically could run at near 100% utilization, how would that impact replacement cycles?

Furthermore, as these vehicles get more and more complex, would you ever want to be in an “aged” unit? 

How would this change your idea of vehicle ownership vs. more on-demand/ shared type of models?

Maintenance operations

Who will be responsible for maintaining the vehicles and which aspects of those vehicles? How will this impact your internal maintenance shops and the technician expertise you’ll need to have on hand?

The Bottom Line ...

So much has happened in the AV space since Google put its first self-driving Prius on the roads--only 8 years ago.

And industry consensus is that we’ll start seeing at least Level 3 capable autonomous vehicles to hit the market legally in 2021—which is only 4 years from now.

And the latest trends appear to be backing this up.

But there are still big hurdles and many questions--beyond the technology-- that government, industry and citizens must answer before robots will rule the roads. 

One thing’s for sure. Your fleet operations—and your career—will be impacted in a big way. 

So, how will you adapt?

That’s the question I leave you with today.

Thank you.

About the Author: Sean M. Lyden is CEO of Lyden Communications LLC, a strategy & storytelling consultancy that applies journalistic storytelling techniques to advise CEOs on what to say (strategy) and how to say it (storytelling) through articles, books, and speeches that move hearts, change minds, and sell big ideas that change the world. He also serves as editor for Utility Fleet Professional magazine and writes about the future of transportation and the impact on business and society.