Fully autonomous vehicles are coming to Tennessee’s roads, bringing many more changes than the absence of a steering wheel. Just as America’s interstate highways changed the way we lived, worked, traveled and made decisions in the 1950s and 60s, the age of self-driving cars should deliver significant societal benefits, including reduced accidents, injuries and fatalities, as well as improved traffic management since cars and trucks will be connected, both to each other and traffic signals.
All of this will impact industries and employment, with one estimate stating self-driving cars could eliminate 3 million U.S. jobs.
Self-driving cars also have the potential to affect where we live and work, perhaps reversing the current trend toward urban living.
Being able to tolerate a longer commute could make living farther from work and downtown attractions more palatable, perhaps even easing affordable housing issues. That hour or two-hour commute might become the best part of the day as riders reclaim their time to watch a movie, reply to emails or take online courses.
“Suddenly that downtime, travel time, becomes potentially productive time,” says Mike Wrye, senior vice president of Lose & Associates, an architecture and engineering company with offices in Nashville and Atlanta.
“With the technology we have today, you could theoretically be in your autonomous vehicle and conducting business. Much of what I do, if I have a phone and a laptop, I can do it. That’s the case for a great number of professions nowadays.”
That easier commute might create entire new communities even farther out from the suburbs of La Vergne or the new apartments cropping up in Gallatin. If people can use travel time for work or entertainment, living near a city center won’t be as critical.
And that could affect the value of rural land, making it more attractive to developers than farmers.
“The proximity to the destination you drive to every day is less important,” Wrye says. “I think it lends itself more to urban sprawl, but frankly development patterns have been going that direction anyway.
“So, the emphasis really has been the last decade or so to try to reverse that. That’s where you see the result of the focus by the national planning commission on trying to emphasize density on town and minimize sprawl. But, it’s still occurring.”
How soon will all this happen? The technology is already here.
A number of today’s cars already can help drivers avoid drifting into adjacent lanes, warn them if there is something behind them when they are backing up or brake automatically when coming upon a sudden stop, thanks to a combination of hardware and software that assists drivers.
The National Highway Traffic Safety Administration is already preparing for the time fully autonomous cars and trucks will be a reality, predicting fully-automated safety features and highway autopilot will be available sometime after 2025.
Jim Hines, an analyst with Gartner, a research and advisory company, says there are key challenges as society moves toward fully autonomous vehicles such as cost reductions for the technology, and legal and societal considerations such as liability, insurance and what it means for human interaction.
Wrye, whose job requires anticipating the future, says changes in lifestyle could potentially take decades as autonomous vehicles become widely successful.
“The problem with the people I talk to, and I see, is that even though the vehicle itself may function perfectly, there’s going to be a transition period,” Wrye explains.
“And until you have the vast majority of drivers not driving, simply riding, then potential benefit for autonomous vehicles really doesn’t materialize.”
Traffic, parking
Partially autonomous vehicles will likely be on the roads by 2020 or 2021, and people in transit positions across the state are trying to determine how they will fit into traffic and infrastructure plans in the next decade.
Erin Hafkenschiel, director of the Nashville Mayor’s Office of Transportation and Sustainability, says it is difficult to determine exactly how autonomous cars will change traffic, but from the data that her office is seeing, autonomous vehicles won’t solve congestion on their own.
“Some of the data that we’re seeing out of the Colorado Department of Transportation is saying that they could potentially take your vehicle lane capacity from 1,600 vehicles an hour to 2,000, maybe 2,200, vehicles an hour, whereas a four-car light rail train can move 16,000 people per hour,” she says.
That’s why Nashville’s most congested corridors need some type of high capacity transit, whether it’s rapid bus or light rail, to help increase the capacity of the roadways.
In Nashville, autonomous vehicles are a small part of the current $5.2-billion-dollar plan being voted on in May. Hafkenschiel says there’s just no way the city can put all its eggs into the self-driving basket, and that the proposed plan is as state-of-the-art as possible.
“Nashville is growing really fast,” she says. “We have a million more people that are moving to the region in the next 25 years. And if those million more people bring a million more cars, then our whole region will come to a halt.
“And even if you do have this new technology where you can summon a car to your front door at even less money than it costs you to do that with an Uber or Lyft driver today, we don’t think it’ll solve congestion.”
Hafkenschiel agrees autonomous vehicles could have other unintended consequences related to urban sprawl and existing farm land.
“The way they’re being portrayed is that once we have autonomous vehicles, and there is no driver involved, that the demand for them will be sort of limitless. That no one in their right mind would choose walking to a train stop and getting on a light rail train when you can just request a self-driving car to come pick you up from your door,” Hafkenschiel points out.
“You can imagine people moving even farther outside of Davidson County, and farther into the region and just driving in their car, and they’re doing work or they’re doing their makeup or they’re eating breakfast. So, it doesn’t really bother them that they’re stuck in traffic and they’re commuting for an hour and a half each way.”
She says the proposed transit plan addresses the coming world of the self-driving car by allowing the city to grow in a more sustainable, condensed way.
“We want to preserve our rural land, our farmland,” Hafkenschiel adds. “We want to grow around downtown and midtown and along our transit corridors. So, let’s put transit on those same corridors and then let’s do transit-oriented development districts so that more of those million more people that are coming will actually move to be adjacent to these corridors.”
In 1960, people between ages 35 and 65 who traditionally wanted a larger home on a larger lot size with a multi-car garage accounted for 70 percent of the home buyers, Hafkenschiel says. By 2040, that number will fall to less than 30 percent.
Planners consider new home buyers – millennials and the downsizers 65 and older –more likely to live in urban core neighborhoods and walkable transit-oriented locations.
“So, we’re actually just planning to meet that demand, not even trying to change the habits of those people in the middle,” Hafkenschiel says. “If you live in Belle Meade and your house is on an acre lot and you drive into work every day, I actually don’t need you to get out of your car and ride the system for it to be successful.”
‘Ghost cars’
And if people begin to move farther out, there might be even more need for downtown parking, not less.
“It is kind of a reach, but if that mode of transportation becomes the norm and it’s not as undesirable to live and have to travel a great distance to get to work, then you would assume there will be less motivation to carpool or look for other modes of transportation because you can accomplish in your vehicle everything you can,” Wrye says.
“More than likely people would probably be more inclined to travel alone, so you may even need more parking.”
Hafkenschiel says there’s a good chance we will need fewer spaces, not more, particularly if there is a shift from car ownership to a shared-mobility model like Lyft or Uber.
But that also comes at a cost.
“What if you had your privately owned car and it would drive you downtown and then you say, ‘Okay, now go find a parking spot, but I only want you to park if you can find one that’s less than five dollars for the whole day.’ And then that car drove around the city for hours and hours and hours and didn’t stop until it found a $5 parking spot,” she explains.
“So, you have this fear of what people call ‘ghost cars,’ a bunch of cars that are driving around empty.”
Affordability and change
Dr. Tim Lynch, who works in artificial intelligence as president of artisanal computer company Psychsoftpc, says self-driving vehicles will brings changes that might not be affordable for all people.
“For decades, driving a car has been a rite of passage in the U.S.,” Lynch acknowledges. “This will be taken away. We would turn from a society of drive-thrus to one where we have to call for a robocar every time we want to go somewhere, and this can get expensive.
“What the Googles of the world don’t understand is that most folks can’t afford to pay $12 or more for a round trip to McDonalds or Starbucks.”
And then there’s the waiting.
“Today if we want to go somewhere we hop in our car and go,” Lynch says. “If we have to call for a self-driving car every time we will have to wait 10 minutes or more for a ride, and that is both ways.”
Self-driving cars will affect almost every industry, including legal, food, auto repair, trucking, driver’s education, insurance and more, he adds.
“If autonomous cars are successful, and there are less accidents, then there is going to be less need for insurance agents,” Lynch says.
Big companies could incorporate their own insurance in their own autonomous cars and handle that in-house. And if an autonomous car gets into an accident with another autonomous car, who are you going to sue?
“The AI proponents especially are saying it’s like the industrial revolution, and it’s not,” Lynch points out. “It’s not because industrial revolution was machines that didn’t think. You still needed people to run them. You still needed assembly lines. It lessened the handcrafting that people were doing, the hard labor.
“But as the machines moved in, it still created jobs. The problem with AI is as it moves in, it’s going to take over jobs and it’s not going to create jobs.”
More pollution or less?
Rich Davies, deputy director of the sustainable transportation program at Oak Ridge National Laboratory, says researchers are interested in how different policies might shape people’s behaviors and influence the overall energy efficiency if you enable people with autonomous vehicles.
And the consensus is still pretty cloudy.
“A study showed this adoption of connected and automated vehicles could cut the energy consumption by as much as 60 percent,” he says. “And, it could almost double it as well, depending on the decisions that people made and the policies that are essentially deployed around it.”
So, more pods could mean even more pollution, not less. A recent study by U.C. Davis transportation researchers states services like Lyft and Uber are actually adding to congestion and pollution. Plus, users of the ride-hailing services are still keeping their own car.
The study looked at 4,000 users in seven major metro areas – Boston, Chicago, Los Angeles, New York, the San Francisco Bay Area, Seattle, and Washington, D.C. – between 2014 and 2016 and recommended that if cities are looking to tame traffic and carbon emissions they should prioritize high-capacity vehicles like buses over single-occupancy cars, and move to priority lanes.
That’s the plan in Knoxville, where traffic engineer Jeff Branham says the city was recently received awarded a $6.4 million of CMAQ (congestion, mitigation and air quality improvement) funding, a TDOT grant program that funds infrastructure projects that reduce pollution.
That money will go to an Advanced Traffic Management System for traffic synchronization, as well as plans to create Accelerated Bus Corridors on Knoxville’s busiest and most congested arterial roads, like Broadway (U.S. 441), designed to help unclog congestion.
“What we see in the future is a connected vehicle, and a connected user,” Branham says. “Our focus is preparing for that connectiveness that’s needed to provide the services that someone with a mobile app needs to park their car, or a mobile app that is maybe integrated with the busing system so they can understand what bus they need to take.”
Information like whether a pedestrian is getting ready to cross the street or what direction they’re traveling can be shared, Banham adds. Also, as a vehicle enters a work zone, messages about detours and special areas of concern are given.
Part of this project is to create 12 enhanced bus stops, as well, with built-in passenger information systems so that you don’t have to have a smartphone or a computer to get real-time information about how many minutes away the bus is.
The project will also construct special loading zones upstream of six traffic signals so that a bus can get out of the flow of traffic while the bus is loading passengers.
All of this is an effort to ease the added congestion that inevitably comes when a large bus makes frequent stops and has to get back up to speed.
“On Broadway, we have a problem with roadway capacity. Basically, we have a four- or five-lane section, and it’s running at 130 percent of its capacity for about an hour every day, up to two hours every day in certain places,” Branham says. “You see a lot of side swipes, and people getting very frustrated with the busing.”
Not that Knoxville isn’t embracing autonomous vehicles. Olli, a self-driving shuttle made by local company Local Motors, is now in limited-use runs for events and tours downtown and at Chilhowee Park. But Olli won’t be transporting regular daily commuters, so there’s no big effect on routine traffic patterns. But the city is anticipating more self-driving systems in the future.
“A connected vehicle is much more aware of its surroundings, so it tends to be much more efficient, and it tends to avoid traffic conflicts and collisions more than a human driver,” Branham says. “So, we’re looking forward to the safety increases of a more autonomous future.”
Davies says they want to make sure that the energy impacts of different policies can be understood by the people who do actually set them, as well as safety estimates. As a US Department of Energy laboratory, Oak Ridge has a lot of DOE and US Department of Transportation projects looking at ways to develop these different systems and traffic control approaches, to try to automate these things as they’re deployed over the coming decades.
“If you look at the cost every year on the US highway system, the number of fatalities, the amount of injuries, and the overall cost of all the accidents on the US Highway system, basically we fight a Vietnam War every year,” Davies adds. “I would postulate that many if not most of those things could actually be reduced if you have artificial intelligence instead of human drivers, but that’s just my opinion at this point.”
Safety the ultimate goal
Safety is the biggest motivating factor for most proponents of the fully-autonomous driving that will eventually eliminate human drivers. After a steady decline in traffic fatality deaths statewide from 2012 to 2015, the rate jumped to a new high of 1,039 in 2016 thanks to the rise in distracted driving.
Data from the National Highway Traffic Safety Administration points to human error or bad decision contributing to 94 percent of crashes so by removing humans from the driving process completely, many people think it will drastically improve safety on Tennessee’s roads.
Rep. William Lamberth (R-Cottontown) is chair of the House Criminal Justice Committee, and the very thought of autonomous vehicles eradicating drunk driving or distracted driving deaths highly appeals to him.
“I worked for a long time both as an assistant district attorney, as a private attorney, and I have – we all have – lost loved ones and friends to either intoxicated drivers or distracted drivers, and just the thought of that not even being a part of our society anymore is exciting,” he says.
“It would be amazing if we never had to open the paper and read about a family or a father or a son or daughter or husband or wife that’s been killed by a drunk driver or someone that was driving distracted.”
A NHTSA study showed motor vehicle crashes in 2010 cost $242 billion in economic activity, including $57.6 billion in lost workplace productivity, and $594 billion due to loss of life and decreased quality of life due to injuries. Eliminating the vast majority of motor vehicle crashes could radically reduce these costs.
But that safety incentive can come with a cybersecurity cost.
Jared Smith is a cybersecurity researcher at Oak Ridge National Laboratory where he is the principal investigator for several projects with funding in excess of $1.5 million in vehicle security, computer network and internet security, industrial systems and IoT security (internet of things - a network of devices, cars, home appliances, etc. embedded with electronics, software, sensors, actuators, and network connectivity).
Smith’s research group has a dedicated vehicle security center that includes autonomous vehicles, a growing need since Smith says the goal in the next few years is that all new vehicles, autonomous or not, will not only talk to each other, but talk to traffic lights and boxes on the side of the road, exchanging messages in order to get you places quicker.
But those lines of open, mobile communication need to be secure. That’s where Smith’s team comes in, for instance, to ensure anonymity and make sure a hacker couldn’t determine where a specific person was going, intercept the car and hijack the systems.
“We can only do as much as people let us, and right now the automotive industry is pretty restrictive about letting security researchers tinker with their cars,” Smith says.
There are hundreds of people at ORNL working just on vehicle technology, from security to sustainability to energy to 3D printing vehicles and self-driving technology, Smith adds.
“Being open to autonomous vehicles is a big deal, because at the end of the day these things should make our lives a lot safer,” Smith says.
“We already have all the research up here and it’s pushing forward. The real issue is how do we get more funding from the industry to let us do more things with this? That’s something that I know the legislature is actively trying to help us with, and we’re pushing it down our alleys as well.”
Oak Ridge has some of the fastest, high-performance computers available in open science today. They have a strong artificial intelligence and cybersecurity network, and leveraging those capabilities to move into this connected and autonomous world.
Davies says there is a mobility initiative inside of the lab focused on the nexus of connected, automated, and electric vehicles, the three big trends that are going to shape the next 30 years.
“The electric vehicles are important to us, but if you look at what automated vehicles and particularly connected vehicles to do, you can start to control these things in a way that can maximize energy efficiency, reduce fatalities and accidents,” Davies says. “So really that’s the nexus of what we’re looking at.”
Man vs. machine?
Self-driving cars would likely reduce accidents, Lynch says, but it’s not a given. He brings up the man who died in 2016 using Tesla’s Autopilot mode on his Model S, the first fatal crash involving a self-driving car.
“The AI in that Tesla car didn’t know the difference between a semi and a bridge and decided that the semi was a bridge and it could drive under it,” Lynch says. “There’s a concept in AI called perverse instantiation. That is, we give an AI a task like driving. It completes that task, but it completes it in a weird way that we humans would never even consider.”
In that Tesla crash, the National Highway Traffic Safety Administration found no defects in Tesla’s system.
Still, AI just doesn’t have the socialization and people skills humans have, or the life experience to just simply know the difference between a semi and a bridge.
“The AI running the car doesn’t have a moral compass,” Lynch explains. “It doesn’t know right from wrong the way that we do. It just knows there’s a task, and the best logical way to accomplish that task. Sometimes it will accomplish that task in ways that will harm people or harm the environment or harm nature in ways that we humans would never even consider doing. So that’s a problem with artificial intelligence-driven cars.”
Smith is skeptical AI is going to take over the world someday, or that all the cars will sync up and start running people over – at least not on their own.
“If that’s going to happen, it’s going to be somebody that hacked into the system and made them all do that, not them thinking it on their own,” Smith says.
But there are, he admits, some hurdles.
“I think there’s going to be issues with things like ethics behind this,” Smith notes. “What would the car do if you’re going to get in a crash, but you’re one person and the other person is a van of four children and a mom, and it knows that because it knows the people in the vehicle securely because it knows all the people on the road.
“Do you die or do they die? I mean, how do you decide that? Clearly, there’s going to be a lot of government intervention in how these things that are actually modeled.”
So how do you plan for something that requires a Steve Jobs-level of forward thinking?
From insurance to traffic to jobs to safety, self-driving vehicles will change almost every aspect of life.
And businesses, cities and even our own mindset will just have to scramble to keep up.
“We’ll be seeing a lot of changes coming up in all sorts of areas regarding artificial intelligence, and all sorts of issues that we need to be thinking about now before it becomes too late and we can’t do anything about it,” Lynch says.