Driverless rigs are already on roads, but three challenges are slowing expansion
Potential gains from automation
Automated trucks use artificial intelligence, machine learning, cameras and other sensors to assess their surroundings and make driving decisions. Torc, a company developing the vehicles, combines information from multiple sensors to predict how a truck will perform.
Data from big rigs with features such as automatic emergency braking and lane-keeping assistance suggest driverless trucks could reduce crash numbers and severity. They may also use less fuel through optimized driving and close convoys, reduce idling emissions and parking needs, and help address the driver shortage.
Weather and roads pose problems
Rain and snow can interfere with the cameras and sensors automated vehicles rely on to perceive the road. Waymo, a driverless taxi company, operates its cars only in places that do not get snow.
Construction zones can also be difficult when lanes are narrow, cones are present or road markings are missing. The article notes that some research finds human drivers crash into driverless cars more often than into cars driven by people.
Companies narrow their plans
Some developers have reduced their ambitions, changed direction or left the market. Embark, TuSimple and Locomotion closed down in 2023, while Waymo shifted its focus exclusively to cars.
Other companies have concentrated on trucks traveling a small number of dedicated routes between major cities or manufacturing sites. The article’s author expects automated trucks and human drivers to work together for years.
Drivers remain part of the system
Research by Shannon Roberts and her colleagues points to the importance of interfaces that show drivers how automation is performing. Displays could provide updates, while spoken alerts or a vibrating seat could warn when the system cannot manage a situation and the driver needs to take over.
Developers also need to account for automation bias, in which drivers rely too heavily on a truck’s decisions, and algorithmic bias. The latter could cause sensors to recognize some people on the road more readily than others.
Key points
- Automated trucks are being tested on limited routes in Texas and California.
- Rain, snow, difficult road layouts and unpredictable traffic limit reliable operation.
- Several companies exited the market or narrowed their plans in 2023 and afterward.
- Driver alerts and safeguards against bias will matter as humans and automation work together.