Before we can buy cars that drive us to work while we catch up with the news, and before we see robotaxis cruising city streets, we’re likely to see self-driving trucks traveling along highways—possibly as soon as 2025. Why will autonomous trucks arrive first, and why so soon?
When hype for autonomous vehicles (AV) began in 2016, many experts predicted the coming of self-driving cars by 2020. This timeline elongated as companies slowly realized the complexity and enormity of that goal. Traffic, variable road conditions, adverse weather, and the unpredictability of human driving behavior all make driving very difficult for computers.
Highways, however, offer a simpler environment for self-driving trucks to negotiate because there are a limited number of actions required. Autonomous trucks need only merge onto the highway, monitor their speed, change lanes, stay mindful of the cars around them, and exit. These tasks aren’t easy, but autonomous trucks can manage them with relative ease—and even drive more safely than any human could.
Merging and changing lanes: dangerous movesA fully-loaded tractor-trailer truck can weigh up to 80,000 pounds—20 to 30 times more than a car—and might need more than twice the distance to stop. In a collision, a truck’s kinetic energy is often deadly. According to a 2020 Federal Motor Carrier Safety Administration (FMCSA) annual report, there were approximately 415,000 reported truck crashes in 2020, and more than 4,000 were fatal. Many of these accidents can be attributed to two dangerous maneuvers: merging onto highways and changing lanes.
Merging demands that drivers simultaneously observe on-ramps and freeway lanes, find a gap to merge into, and synchronize their speed. Lane changes take approximately 10 seconds, and in those few moments a motorcycle might rapidly approach from behind, or a car might try to squeeze by.
To merge or change lanes safely, AV systems must clearly perceive what is directly in front of and behind the truck, plus any impediments that might be farther away.
New technology prevents collisions
Advanced technology helps prevent accidents for human and autonomous drivers alike. For instance, automatic emergency braking systems reduce the rate of front-to-rear crashes by 41 percent. To assist with rear-facing and forward-facing crash avoidance, new technologies that promise similar effectiveness are being developed to help trucks “look” hundreds of meters ahead or behind them on each side.
Stereo vision systems, for instance, provide long-range, high-density 3D data in real time. A stereo vision system can detect objects as far as 1,000 meters away, and as small as 10 centimeters in height. Considering that trucks need up to 350 meters to stop or avoid obstacles, long-range perception is crucial. Imagine a self-driving truck moving at 65 miles per hour toward a motorcyclist disabled in the roadway ahead. An automated system can detect both the motorcycle and rider, perceive them accurately, and safely avoid a collision.
Other camera-based technologies support rear-facing perception, too, but within limits. Monocular cameras utilizing AI work well but at shorter ranges and with far less reliability than stereo vision. Lidar can be used also, but at shorter ranges, 10 to 20 times lower resolution, and exceedingly high cost.
Beyond human errorTrucking is not an easy life. Drivers can rack up to 3,000 miles a week behind the wheel, fighting bad weather and heavy traffic—along with their own fatigue and sleep deprivation, distraction, and boredom. Sometimes the result is reckless driving or substance abuse. “A single momentary lapse can result in cataclysmic damage,” one blogger writes. Humans are fallible, and human error causes accidents.
Computers, however, work “tirelessly” without such problems. They process input from multiple sensors simultaneously, have better memories than humans, and can be programmed to assess many situations and combinations of situations almost instantly. Even at the most stressful moments, self-driving trucks can make decisions quickly and react faster than we could.
Another plus: While human truck drivers can legally drive only 8 hours a day, autonomous trucks can drive for 24 hours, allowing much greater trucking efficiency. This shift is also likely to reduce day-time congestion and improve traffic flow in cities, with fewer unsettling encounters between urban commuters and 18 wheelers.
All in all then, humans might be happy to share the highway with truck “drivers” that never fall asleep at the wheel, stay in their lane, and always watch the road. It’s possible that in just a few years autonomous trucks will be making the roads safer for us all.