Are Self-Driving Cars in Our Near Future?
Self-driving cars might be the focus of countless news stories these days, but experts are split on how mainstream the autonomous vehicles will be, and how quickly we’ll get there. Some experts predict it’ll only be a few years before the general public makes the switch to self-driving automobiles, while cars that require drivers go the way of the dinosaurs. Others aren’t so sure. They point to gaps in technology, lack of infrastructure, and safety issues as signs we aren’t as close to driverless cars as we might think.
For some drivers, that’s good news. Not every motorist is eager to embrace a self-driving vehicle. There are, however, benefits to autonomous vehicles that can’t be eliminated as long as humans are in control of cars. The question remains: how close are we to the widespread use of self-driving cars?
Understanding the Levels of Autonomous Vehicles
To understand what the experts are talking about when they talk about autonomous vehicles, you need to know what the autonomous vehicle levels are. There are six levels of self-driving vehicles, classed by the Society of Automotive Engineers (SAE) from Level 0 to Level 5.
Level 0 (no automation): There are no automated assistance technologies (traditional cruise control that keeps the vehicle driving at one set speed or beeps that warn about a potential crash do not count towards autonomy, for example). In a Level 0 car, the human is responsible for all driving actions.
Level 1 (driver assistance): The vehicle has at least one advanced driver-assistance feature. This can include adaptive cruise control or lane-keeping technology. The driver supervises most functions, but the car assists in some. Many passenger vehicles fall into this category.
Level 2 (partial automation): The vehicle’s features include two or more advanced driver assistance systems. These systems are responsible for areas such as steering, acceleration, or braking. The technologies work together to keep the vehicle moving, but a human driver is still actively involved in monitoring the vehicle.
Level 3 (conditional automation): The vehicle can take full control of driving during some parts of a trip under certain conditions. For example, a car that can drive itself on the freeway without human intervention (except when conditions require intervention, such as to avoid an accident).
Level 4 (high automation): The vehicle is in control of the entire journey without intervention but has constraints such as being limited to a specific geographical area (for example, a city). A human may need to take control in some circumstances. There are no Level 4 vehicles available for consumers.
Level 5 (full automation): The vehicle is in control under all circumstances, to the point where there are no steering wheels, pedals or joysticks for a driver to control. The vehicle also has no constraints to its abilities — it isn’t limited by geographical boundaries or kept to certain speeds.
Level 1 and 2 are the most widespread in personal vehicles. Most experts agree that commercially available individually owned Level 5 vehicles are a long way off, though some car manufacturers are pushing to make that a reality sooner than later. Autonomous vehicle developers aren’t just focused on privately owned passenger vehicles, however. There are a few areas in which self-driving vehicles can have a positive impact.
When experts talk about self-driving vehicles, they’re typically referring to Levels 3 and 4. Level 3 is more relevant for individually owned passenger vehicles, while Level 4 is more relevant for ride-sharing companies.
What do the Big Manufacturers Say About Autonomous Vehicles?
The companies focused on fully autonomous vehicles tend to fall into two camps: those who are concerned with companies that offer rides to passengers (Uber or Waymo, for example), and those who are developing vehicles that deliver goods. They see those two areas as being the most financially lucrative.
GM, for example, has said it expects to be “the first high-volume auto manufacturer to build fully autonomous vehicles in a mass-production assembly plant.” To that end, the company spent $581 million to purchase Cruise Automation — a self-driving car company — and has announced a new research and development facility dedicated to self-driving cars. GM, however, is focusing on ride-sharing vehicles over individual consumers and has an agreement with Lyft Inc.
Ford is also focusing its efforts on ride-sharing vehicles. The company has said it will have a Level 4 vehicle available in 2021, completely skipping over Level 3 automation. Honda, meanwhile, is in discussions with Waymo and wants to produce vehicles that can drive themselves on highways by 2020. For its part, Waymo plans to launch its ride-hailing service in Phoenix by the end of 2018, making it the company’s first revenue-making service.
Other companies are focused on delivery services using self-driving vehicles. Udelv has a deal to supply deliver vehicles to a large grocery chain in Oklahoma City, while Nuro vehicles will deliver groceries from Kroger to Arizona residents.
What Will Self-Driving Cars do to Car Ownership?
If fully automated self-driving cars become widespread, they could change the face of car ownership. After all, people won’t necessarily have to own a vehicle anymore; they could belong to a self-driving car service and order the car whenever they need it, no driver required. The need for parking could change, too, as a self-driving vehicle could be sent to a different location to park until the next time it was needed, regardless of who owns it. A study conducted in Boston by the World Economic Forum suggested that with the adoption of autonomous vehicles, there would be a 48 percent decrease in the number of parking spaces needed. The number of vehicles on the road would also decrease.
Self-driving shuttle services could also decrease car ownership by working in cooperation with local transit services. Waymo is operating a pilot project in Phoenix to see if it can be used to increase use of public transit. Waymo and Valley Metro are working together to offer a robotaxi to a group of Valley Metro employees who live just outside of walking distance to a Valley Metro stop. A Waymo vehicle shuttles these employees between home and the public transit stop.
The company plans to expand testing to include Valley Metro RideChoice users — seniors and people with disabilities who experience barriers to using public transportation. By making public transportation more accessible, self-driving ride-share companies could have a drastic impact on individual car ownership.
How Far Are We From Self-Driving Cars?
Currently, there are some self-driving cars in geo-fenced areas in a handful of cities. So, if you’re driving in Detroit or Las Vegas, you may have already shared the road with a self-driving vehicle, either shuttling people or products. They are being used with a select group of people and along designated routes for testing.
Most of the vehicles are limited to only a few miles, though as testing progresses the distance they can travel will increase, too. Currently, Drive.ai is operating four minivans as shuttles near Dallas along a two-square-mile area. Soon, though, the vehicles will go to AT&T Stadium.
How far we are from autonomous cars being regulars on city streets with fewer limitations depends on whom you’re talking to and whether they’re focused on privately-owned vehicles or ride-sharing companies. The experts seem to agree that no company is close to producing a genuinely autonomous, Level 5 vehicle. Gill Pratt, CEO of the Toyota Research Institute, said that it is likely a few companies will have Level 4 autonomy cars in certain areas within a decade.
Of course, some experts believe we’re closer than that. Elon Musk says he believes that by the end of 2018, a Tesla will be able to make the full drive from Los Angeles to New York City without a human using the steering wheel.
Changes Still Needed to Get Autonomous Vehicles Going
Despite all the good autonomous vehicles can do — imagine a world with no impaired driving deaths — there are still many changes needed before self-driving vehicles are widely adopted, and trusted. A significant complaint about autonomous vehicles is that they think like machines, not like people. That makes it difficult for autonomous vehicles to predict human behavior and difficult for other drivers to predict the autonomous vehicle’s behavior.
Fast Company gives the example of a self-driving car that refused to move at a green light. The reason it stayed stopped? It recognized a cyclist that was stopped at the traffic light but had stopped just past the white line. The car thought the cyclist was about to move because it didn’t register that the person’s foot was down and he wasn’t in a position to cycle. Human drivers could look at that cyclist and know he was fully stopped and waiting, but the autonomous vehicle didn’t.
Similarly, other drivers have complained about autonomous vehicles stopping during right turns for no reason or slowing down while making left turns. These are all kinks in the system that must be worked out before the public will trust autonomous cars. A company called Perceptive Automata is working on giving machines the ability to replicate human intuition — that intangible quality that we use to determine whether a pedestrian is about to cross the street or whether another driver sees us.
The ability to think like a human is essential for self-driving vehicles. Humans use their intuition constantly. Consider coming upon a vehicle pulled over to the right in a lane of traffic to allow someone to get into the car. A human could see the door open and the person move to get in, and know that they can drive around the stopped vehicle (provided there’s no oncoming traffic). An autonomous vehicle might wait a long time for the stopped vehicle to move again.
Because human drivers might logically expect another human driver to pull out and pass the stopped car, the autonomous vehicle becomes unpredictable and frustrating for other drivers. Therein lies the danger: human drivers predict the movements of other vehicles based on human reactions. If autonomous vehicles behave differently, human drivers can’t predict what the self-driving vehicle will do.
A further factor is the ability of autonomous vehicles to navigate in poor weather. Snow and rain have been shown to affect LiDAR and cameras, while vehicles have difficulty adapting to icy road conditions. So far, much of the autonomous vehicle testing has been done in sunnier climates.
“Snow not only alters the vehicle’s traction but also changes how the vehicle’s cameras and sensors perceive the street,” the World Economic Forum wrote.
Regulations are another critical factor in how soon self-driving vehicles can be widely adopted. There are important legal ramifications to having self-driving vehicles on the road, and regulations must be developed to account for this. It’s likely not a matter of if the legislation changes, but when. Still, delays in implementing legislation or in changing the infrastructure will mean it takes longer before self-driving cars are widespread.
So how long will it be before self-driving vehicles are as common on the road as those requiring drivers? According to Waymo CEO John Krafcik, it will be “longer than you think.”
Sources:
https://www.cnet.com/roadshow/news/self-driving-car-guide-autonomous-explanation/
https://www.techemergence.com/self-driving-car-timeline-themselves-top-11-automakers/
https://www.theverge.com/2018/9/17/17859112/self-driving-cars-shuttle-pods-delivery-services
https://www.fastcompany.com/90231475/perceptive-automata-helps-self-driving-cars-think-like-humans
http://www3.weforum.org/docs/WEF_Reshaping_Urban_Mobility_with_Autonomous_Vehicles_2018.pdf
https://techcrunch.com/2018/09/11/apple-autonomous-vehicle-test-fleet-california-70/
https://thenewdaily.com.au/life/tech/2018/09/03/driverless-cars-experts/
https://www.wired.com/story/aurora-self-driving-cars-plan/
https://www.boston.com/cars/car-news/2018/09/12/self-driving-cars-boston
