Autonomous trucking 101: What is it and where the industry is headed

Autonomous driving, in general—and autonomous trucking, in particular—have had quite a bumpy road so far. While the technology for autonomous driving is no longer in its infancy, there are still a lot of concerns about this field: Is it safe enough? Will it take jobs away from people? Is it just hype with no substance?

We’ll answer all these questions and more. Let’s hit the road!

What is autonomous trucking?

Autonomous trucking refers to self-driving trucks that require no intervention from human drivers to safely operate. Self-driving trucks combine various sensors and AI-based software to control and drive the vehicle.

Currently, there aren’t any fully “driverless” trucks operating in the U.S. There are, however, a lot of cars and trucks with varying amounts of automation. Adaptive cruise, lane and brake assistance are the three most common features in semi-automated trucks.

How does autonomous trucking help?

It’s no secret that we’ve been facing a multi-year shortage of qualified drivers for over-the-road truckers. Two years ago, the American Trucking Associations (ATA) estimated the driver shortage at 60,000 in the over-the-road truckload market. This shortage is not likely to stop anytime soon. Thus, autonomous trucking is not here to steal jobs but to supplement where skills shortages are most severe.  One of the trucks we outfitted recently shaved 10 hours off a 900 mile run.  On average, we are seeing 42% faster transport because of the reduction in required breaks.  This is all great news for the consumer.

Safety is another concern that autonomous driving helps alleviate. Everybody’s familiar with drivers who accelerate quickly. Everybody’s familiar with drivers who brake abruptly. Humans are inherently fallible, whether or not we’re willing to admit it.

Autonomous driving solutions can help mitigate some safety concerns by improving predictability in how the vehicle performs on the road; smoothing out inconsistent acceleration, speed and braking; and overall increasing the efficiency of the vehicle. These controls increase the reliability of the truck, as well as the fuel efficiency, and allows operators to accomplish more with their vehicle.

Artificial intelligence—or AI and machine learning—collects, interprets and prioritizes data. It renders full 360-degree, real-time images and conditions, including visibility—many times farther than the human eye can see, even in the dark or in the rain.

More importantly, the autonomous system doesn’t get distracted like a human driver. A computer isn’t concerned with changing what’s on the radio, having a conversation, texting, drinking a cup of coffee—and better yet, the computer has no blind spots.

What are the levels of driving autonomy?

There are five levels to driving autonomy. Level 0, or no automation, is one-hundred percent human operation. You probably can’t find a car on the road today that is at this level anymore.

Most passenger vehicles on the road have level 1 automation, which is driver assistance that offers technology such as cruise control, adaptive cruise control and lane assist.

Many of today’s newer vehicles are coming out with level 2 or partial automation, which includes self-parking, automatic braking and even lane-changing. At both Level 1 and Level 2, the driver is still hands-on and responsible for the overall operation of the vehicle.

When we get to level 3, or conditional driving automation, this is where the vehicle can accelerate around a slow-moving vehicle; monitor its surroundings; change lanes; and control the steering, throttle and braking. All the driver has to do is pay attention and be ready to take back control when necessary.

When we get to level 4, or high driving automation, this is where the human interaction with the machine is greatly reduced as the vehicle’s capacity increases. Steering, braking, accelerating and monitoring the environment are taken out of the driver’s hands, as are changing lanes, signaling and turning.

Level 5, or full driving automation, requires zero human attention. There’s no need for a fully autonomous vehicle to have a steering wheel, brakes or even a gas pedal. The vehicle controls all driving tasks under all conditions, including monitoring the environment and identifying complex driving conditions like busy pedestrian crossings.

That’s the general scale of vehicle automation. Where are autonomous trucks on this scale?

They are currently between level 2 and level 3. Equipped trucks are capable of automated operation but still require the driver to be responsible to respond or intervene if necessary. Current autonomous systems offer comprehensive situational awareness with advanced technology such as radar, lidar and video cameras, which together provide massive amounts of sensor data to the brains of the system—which, of course, is the server.

As of now, autonomous trucking comes down to a partnership or a balance between the driver, who can execute abstract thinking that computers can’t fully do yet, and automated systems and sensors that can see and process what the driver can’t.

How does rugged computing fit into autonomous trucking?

Let’s look at an analogy that’s easy to understand: Human brains are very good at abstract thinking and computers are very good at linear problem-solving. Give a computer a very specific problem to solve, it will do so quickly. Ask a computer to be creative, you’re often disappointed with the results.

When you present a computer with unknown conditions, it doesn’t always respond the way a human would. To approximate how a human would direct a vehicle in surprising circumstances when things are unpredictable out on the road, it requires an incredible amount of processing power. And even advanced AI systems can only come so close to approximating the amount of abstract thinking that a human being can do.

Unique conditions and requirements of vehicles are driving the autonomous driving market toward rugged computing solutions. The market started off by using standard off-the-shelf computers. And very quickly, developers and integrators discovered that the systems were not up to handling the shock and vibration of going over North American highways. You can joke about the potholes in your particular state all you want, but every state has roads that are not computer-friendly.

They also quickly discovered that the technology was cooking itself. The amount of computing power for autonomous driving generates a lot of heat—about the amount of two space heaters. We see computer heat loads running in the 1½ to 2 kilowatt (2,000 watts) of heat being generated by the systems that are driving vehicles down the road.

Put the equivalent of two 1,000-watt space heaters in the rear of a vehicle and the hardware would bake itself to death. Put it up in the cabin of the vehicle with a human, the safety operator, it’s going to get very, very uncomfortable for them. Removing heat from the vehicle—away from the operators, away from the technology and out into the ambient air while producing little noise—is a critical piece of the puzzle to making autonomous trucking a reality.

A rugged solution that can handle harsh environments, can handle elevated temperatures and can remove heat quickly and efficiently is crucial to keeping these systems operating reliably and comfortably while they’re over the road.

Crystal Group has developed several different approaches to meeting these thermal requirements. An air-cooled or liquid-cooled solution paired with our rugged servers can be deployed into a vehicle to allow high-performance computing components to operate without jeopardizing the driver or sacrificing performance and reliability.

Computing systems need to be reliable to maintain critical data processing and decision-making being performed by autonomous driving software in the vehicle. Crystal Group’s experience in applying mil-spec standards means Crystal Group’s RIA and other rugged computer devices deliver seamless performance under harsh conditions.

What’s next for autonomous driving?

Autonomous vehicles are a developing, growing market. It’s a very exciting place to be working on technology because the requirements change regularly. The problems that autonomous vehicle developers are trying to solve are migrating, mutating and changing regularly. There is no single, one-size-fits-all solution that can meet every autonomous vehicle’s computing requirements. There’s no magic bullet for this industry.

Thus, manufacturers, vendors and solution providers are starting to come up with specific solutions for each industry. For instance, Crystal Group is focusing on autonomous trucking, autonomous buses and autonomous construction equipment—developing specific solutions for markets where our products can be best applied. Doing so allows us to divide the big problem up into bite-sized pieces, into manageable technology solutions that can adapt as the industry grows.

Want more information about our work in autonomous trucking and how it can help your business? Reach out.