Self-Driving Cars Explained - Union of Concerned Scientists

How they work

Various self-driving technologies have been developed by Google (Waymo), Uber, Tesla, Volkswagen, and other major automakers, researchers, and technology companies. They build on decades of advanced driver-assistance systems development, such as cruise control, collision warnings, and lane assist.

For more information, please visit our website.

While design details vary, most self-driving systems create and maintain an internal map of their surroundings, based on a wide array of sensors, like radar. Waymo’s vehicles use cameras, sensors, Light Detection and Ranging (lidar), radar, and highly-detailed maps to get around. AI is used to continue to improve their performance.

Software then processes those inputs, plots a path, and sends instructions to the vehicle’s “actuators,” which control acceleration, braking, and steering. Hard-coded rules, obstacle avoidance algorithms, predictive modeling, and object detection helps the software follow traffic rules and navigate obstacles. Self-driving cars can also be remotely operated or directed by human intervention using teleoperation to resolve unusual or ambiguous situations. Self-driving cars could also potentially communicate with other vehicles and/or infrastructure, such as next generation traffic lights if the technology is widely deployed. Most vehicles in operation today do not currently rely on this capability.

Impacts

The costs and benefits of self-driving cars are still largely hypothetical. More information is needed to fully assess how they’ll impact drivers, the economy, equity, and environmental and public health.

Safety is an overarching concern. Around forty thousand people die in motor vehicle crashes every year in the United States. Self-driving vehicles could increase the miles driven on our roadways —coming with increased safety risks — or, they could improve safety if software proves to be less error-prone than humans. Regardless of what the future brings, we need regulations in place to ensure companies operate responsibly and prudently as testing occurs on public roadways.

Equity is another major consideration. Current self-driving vehicles are developed in private sector applications with no requirements or regulations to serve people that need them most. In addition, there are inequities potentially baked into the technology. For example, studies have shown that automated vehicles are less able to detect people of color and children. Widespread adoption of autonomous vehicles could displace millions of people employed as drivers, negatively impact public transportation funding, and perpetuate the current transportation system’s injustices.

Environmental impacts are a serious concern, and a major uncertainty. Accessible, affordable, and convenient AVs could increase the total number of miles driven each year. Increasing vehicle usage will cause emissions to increase. If these vehicles are powered by gasoline, then transportation-related climate emissions could skyrocket. However, if the vehicles are electrified—and paired with a clean electricity grid—then the increase in transportation emissions could be more modest. If electric AVs replace miles that would be otherwise travelled using conventional gasoline vehicles, it is possible that emissions would decrease.

Emissions could drop compared to our current system, to the extent that electrified AVs enable more shared rides or greater use of public transit for parts of trips.

Seven Principles for AVs

The Union of Concerned Scientists has developed a set of seven principles to help guide how policymakers, companies, and other stakeholders approach this transformative technology. The principles are intended to help maximize the benefits of self-driving cars—and minimize the risks.

Principle #1: Safer transportation

Self-driving cars could reduce vehicle-related fatalities, as machines may make fewer errors than humans—but it’s not a guaranteed outcome. Policy must improve the safety of all road users, whether they are driving, walking, or biking.

Principle #2: Cleaner vehicles

By making transportation more convenient, self-driving vehicles could increase the number of miles that people travel, creating more vehicle pollution. To avoid this outcome, policy must ensure that self-driving technology is paired with clean, low-carbon vehicles, such as plug-in electrics. Policy must also incentivize ride hailing services like Uber or Lyft to maximize the number of passengers for each trip and minimize trips with no passengers.

Principle #3: Integrated transit

AVs will make transportation easier for many—but they shouldn’t replace mass transit options like buses or subways. Instead, AVs should connect transit hubs, provide public transit services to communities not currently served, and generally be used to improve public transportation.

Principle #4: Improved access

In its current form, the US transportation system fails to serve all communities equally, with disadvantages arising based on income, age, race, disability, and geography. Policy should ensure that the widespread adoption of self-driving cars doesn’t perpetuate or worsen existing inequalities and instead gives all people access to clean, affordable transportation options.

Principle #5: Just transition

Self-driving technology will create jobs for some and reduce employment opportunities for others, especially in the trucking and ride-hailing industries. Policy should support career pathways and transitions for affected individuals, and ensure that new jobs are made available to underrepresented populations.

Principle #6: Public accountability and data security

Whether self-driving vehicles improve public health, decrease traffic congestion, or reduce climate change will depend on informed, science-based policy. This will require a robust research agenda and accessible data on the performance and operation of self-driving vehicles. Policy must facilitate open data-sharing, while ensuring that appropriate privacy and cybersecurity protections are in place.

Principle #7: Livable communities

Autonomous vehicles could exacerbate sprawl, congestion, and pollution— or they could lead to more shared rides and fewer cars, freeing up public space. Policy should prioritize the needs of the whole community, not just individual vehicles.

At Torc, we get a lot of questions about our trucks. Whether we’re visiting our local high schools or answering questions via social media, it’s clear that there is a lot of curiosity (and trepidation) about autonomous semi-trucks – and self-driving technology itself. Because we’re passionate about bringing education to the forefront of the conversation about autonomy, we’re taking the plunge into some of our most frequently asked questions.

What are autonomous trucks, or automated semi-trucks?

An autonomous truck is a semi-truck that can deliver goods across the country without needing a driver. Like autonomous cars, these trucks are currently in development at several self-driving-technology companies – and they’re on the horizon with some truly amazing capabilities.

It’s important to remember that all driverless technology exists on a spectrum. Passenger cars and other vehicles with self-driving technology are all around us; you may even drive one yourself! If your car has cruise control, you have a car with self-driving functions – even more so if your car has lane-keep-assist features to help keep you between the lines.

However, at Torc, when we talk about autonomous semi-trucks, we’re talking about full self-driving capabilities. These commercial vehicles use a combination of hardware, software, artificial intelligence, machine learning, and more, to haul freight down highways without needing a driver at some point in the future. Because of their capabilities in speeding up shipping and increasing overall efficiency in the freight industry, many companies including Torc believe that self-driving semi-trucks will be the first commercially viable, autonomous vehicles to scale widely after their initial launch.

With this in mind, there are a few different business models for automated semi-trucks: hub-to-hub and platooning are some of the most popular.

In platooning, sometimes called convoy trucking, one truck follows another, and the lead truck is being driven by an engaged, CDL-licensed human driver. The secondary truck that follows also has a human driver, but that driver is off-duty. (The amount of time that truckers are allowed to drive per 24-hour -period are limited by “hours of service” regulations.) While the second driver sleeps and/or rests, automated trucking software, or a virtual driver, takes over. This virtual driver uses the truck in front of it as a signaler for many of its processes, including braking, accelerating, and even lane-changing.

At face value, platooning may seem like a good path for autonomous trucks, because a human is still involved/in the truck in case something unexpected happens. Secondarily, there are already federal laws set up around team driving; the platooning model just turns one of those team members into a robot.

However, there are some parts of this model that are potentially problematic. In the platooning model, unexpected variables are the enemy. If the follower truck deviates in any way from the lead truck’s behavior or characteristics, things can get dicey. For instance, different set of tires might mean that one truck takes more or less time to brake compared to its leader and vice versa. In the same vein, the trucks may be carrying different loads, meaning that the weight distribution will be vastly different between them, also affecting things like acceleration and braking time.

There’s also the insurance risk. While liability is a hot topic in all forms of self-driving technology, it’s significant in the platooning model. A technological failure in one of the platooning trucks may result in catastrophic damage to both vehicles, creating twice the damage compared to a single, solitary, self-driven vehicle.

And lastly, while there are federal laws surrounding team-based driving, individual state law varies on the platooning concept. Some states, like Arizona, Texas, and Utah, allow automated-follower platooning. However, many other states have “following too close” laws in place that may pose a barrier.

rino.ai are exported all over the world and different industries with quality first. Our belief is to provide our customers with more and better high value-added products. Let's create a better future together.

Now, let’s discuss hub-to-hub trucking automation. At Torc, we believe that the hub-to-hub model is the safest, most scalable form of autonomous trucking. We’ve built our business around this concept, as have many of our advisors, partners, and colleagues. From our freight pilot with Schneider to our collaboration with Daimler Truck North America, Torc has built our hub-to-hub model on the proven value of our partnerships.

So, what exactly is the hub-to-hub model? In this approach, human drivers haul semi-trailers on non-highway roads (think city streets and in-town roads, etc. where drivers encounter things like street lights, stop signs, pedestrian traffic, etc.). In the “first mile” of the hub model, human drivers pick up trailers from places like warehouses, depots, big-box stores and other locations to deliver to highway hubs. These highway hubs will act as “home base” for autonomous trucking, and would be built specifically to manage autonomous truck needs. Once that human driver drops off a trailer, an autonomous semi-truck will pick up that same trailer and haul it across “the middle mile:” those long stretches of interstate where over 8,500 trucks travel every day.  The freight will be autonomously driven on interstates and highways to the next hub, where that trailer is picked up again by a human driver for in-town deliveries, or “the last mile.”

The hub-to-hub shipping concept isn’t new. Most postal and freight services use some form of hubs; you may have even seen them as stops along the route when you track a package online. Whether that’s a warehouse, an office, or a distribution center, many of our daily packages as consumers go to logistics centers where they can be organized and transported to their next stage of transit. As we bring this new technology into reality, we anticipate that autonomous 18-wheelers will integrate seamlessly within an existing shipping hub’s logistics systems and daily operations.

“The benefit of the hub-to-hub model largely lies in complexity and efficiency of the journey,” says Adam Shoemaker, Torc’s Chief Autonomy Architect. “The US highway system, while still highly variable, provides a reasonably structured environment which enables strategic breakdown and development of efficient and safe vehicle functionality. Similarly, the limited interactions of autonomous trucks with pedestrians, bicycles, and other unstructured use cases within a hub model give the product the flexibility to ensure safety without compromising efficiency and marketability.”

It’s also important to note that self-driving “last mile” delivery trucks can be autonomous, and fall within their own category of autonomous technology.  Most autonomous trucking companies are focused on some form of long-haul (middle mile), wherein last-mile delivery is handled by smaller, dedicated robots: think aerial delivery drones and small ground-based robots.

“We believe that autonomous semi-trucks will be the first to make it to market with full autonomy and scalability.”

Adam Shoemaker, Torc’s Chief Autonomy Architect.

What will be the first self-driving vehicle to make it to market?

It can be argued that self-driving vehicles have already made it to market: airplanes operate on autopilot, autonomous cargo ships are on test voyages without a crew, and consumer cars can be purchased with significant degrees of self-driving programmed in. However, none of these vehicles are ready to operate completely on their own just yet.

“We believe that autonomous semi-trucks will be the first vehicles to market with full autonomy and scalability,” Adam noted. “We’re a software company with deep roots in this space as autonomy pioneers, ranging from our DARPA win, to our previous successful autonomous vehicles for defense and mining, and our ongoing collaboration with Daimler Truck and the Torc Autonomous Advisory Council.”

Torc’s expertise in creating safe and dependable autonomous vehicles is proven. Now, we’re proving our trucking proficiency with initiatives like our freight pilot with Schneider. As we continue to move freight for customers, heighten truck perception, and grow our team, we’re racing towards autonomous trucks: the first scalable, sustainable, and practical self-driving vehicle.

Autonomous trucks might seem like something out of science fiction – especially considering how big and heavy the average 18-wheeler is. In comparison to self-driving cars, creating autonomous semi-trucks sounds impossible. But with years of strategy and careful planning behind them, driverless trucks are both possible and probable.

Most newer cars on the road already have some form of driverless technology built into them. Cruise control, lane-keep-assist programs, and more are all examples of everyday autonomy. Although it may seem like we’re a long way from any form of safe self-driving technology, we already live, work, and drive with this new innovation.

However, when most of us think about self-driving cars, we likely think about the kinds of cars that can completely drive themselves. As we’ve seen on the news and likely read about on our social media feeds, fully self-driving cars have a long way to come before they’re completely capable of operating autonomously.

Many of the issues we see in the news about fully self-driving cars, such as stalling in the middle of a roadway or on-road collisions, are a result of an over-reliance on environmental perception with just cameras. But cameras aren’t able to identify things like depth perception, meaning that they can often make more mistakes than the kinds of sensor suites that are on automated big-rigs like Torc’s. And because those sensor suites require significant financial investment, it’s unlikely that we’ll see them added onto any consumer vehicles hitting our city streets anytime soon.

That said, autonomous semi-trucks offer a new set of possibilities for fleet managers and freight professionals across the country. Increasing fleet utilization, shipping loads faster, and improving driver satisfaction make the investment well worth the while.

When will automated semi-trucks be on the road?

The short answer is that automated semi-trucks are already on the road.  In Torc’s case, our trucks drive themselves on highways around Virginia, New Mexico, Texas and Arizona, but it’s important to note that, as of today, there is always an experienced, CDL-licensed human “safety driver” in the driver’s seat, with his or her hands hovering close to the steering wheel, ready to take over in the event that the truck does not behave as expected or unanticipated roadway activity like construction, bad weather or emergency vehicles.   While self-driving trucks are on public roads in a testing capacity, they’re not quite ready to roll out without a driver and safety conductor in the cab. In the vast world of driverless technology, there’s still a lot of development to be done.

At Torc, our completely self-driving 18-wheelers (with no human driver in the seat) will be offered commercially when it is safe to do so. Our primary focus now and always is ensuring that these vehicles are as safe as possible for everyone who interacts with them on roads, at hubs, and everywhere these semi-trucks will go.

And although safety is at the forefront of autonomous vehicle technology, there are other aspects to consider when we think about the product timeline. Topics like insurance and government regulations, as well as certain truck capabilities are still being evaluated in the greater autonomous plan. As an industry, we make leaps and bounds towards answering these questions each and every day, but there’s still a lot to figure out.

At our self-driving technology company, we’re working with our partners, fellow industry professionals, and government officials to bring this technology forward together. In order to successfully launch and scale this life-saving software, we have to create a product that works – and excels – across all levels of stakeholders.

Where are driverless trucks being used?

The latest iterations of the Torc self-driving Freightliner Cascadia

Driverless trucks are currently being used in testing and commercial capacities by autonomous technology companies across many parts of the U.S.   In many states like New Mexico, Arizona, Virginia, Texas, Oklahoma and California, self-driving eighteen-wheelers are testing and delivering goods on public highways.

RELATED: Which States Allow Autonomous Driving?

While we’re not operaing our automated semi-trucks without human drivers yet, we are working towards a future where our vehicles can operate without a human driver on highways. With safety at the forefront of everything we do, we envision a future where trucks can carry freight to and from shipping hubs just alongside the interstate.

Are self-driving trucks the future?

Self-driving 18-wheelers have the ability to make our lives easier in countless ways. From lowering the cost of shipping to creating better jobs for transportation and freight professionals to helping trucking companies operate their fleets more efficiently, this new technology is going to carry our supply chains to new heights of reliability.

And as we work together to create the first generation of self-driven trucks, we thank our Torc team members for lending their expertise to this article. We’d also like to thank our autonomous community for the following sources used in the creation of this article:

• Understanding Perception and Motion Planning for Autonomous Driving, Betty LD. July 8th,
• A Survey of Deep Learning Techniques for Autonomous Driving, Elektrobit. March 24th,

Many of these questions were also provided by PAVE, or Partners for Automated Vehicle Education. Whether it’s self-driving semi-trucks or safety and software, our membership with PAVE allows us to tackle the information challenge in new and innovative ways.

If you’d like to follow our journey to design safe, scalable self-driving 18-wheelers, follow us on social media and check our newsroom regularly for updates.

For more information, please visit Autonoumous Driving Cargo Van.