Semiautomated truck caravans could be plying American highways by 2020—and completely unmanned trucks could join them by 2025—if technology continues to evolve at current rates, according to a study by the consulting firm Frost & Sullivan.

These driver-assisted—or eventually driverless—trucks could help logistics and fleet-management professionals deliver valuable loads with lower costs and improved safety, thanks to faster reaction time for collision-avoidance situations, the firm said.

However, planners face several challenges first, such as installing sensors and networks into a "smart driving environment" and developing reliable vehicle-to-vehicle and vehicle-to-infrastructure communications, according to the report titled "The Future of Mobile Robots in Logistics."

"All of this is a possible reality because we are seeing a lot of development in deploying intelligent traffic management systems," Archana Vidyasekar, the team lead for Frost & Sullivan's Visionary Innovation Research Group, said in a webcast last week. "We're seeing cities and urban areas taking concrete steps toward creating some sort of essential operating platform to control every aspect of the city, especially traffic and vehicles."

The foundation for autonomous trucks is being laid in warehouses as logistics companies find increasing applications for robots, she said. Current examples include the automated guided vehicles (AGVs) seen following preset magnetic tracks through many warehouses and the bright-orange drive units made for Amazon.com Inc.'s distribution centers by Kiva Systems, the company Amazon acquired in 2012 for $775 million and recently rebranded as Amazon Robotics.

Most current autonomous truck applications operate at the semiautomated grade, the first of three automation levels, Vidyasekar said. This "level one" semiautomation relies on familiar 21st-century automotive technology such as keyless entry, advanced driver assistance systems (ADAS), electric power steering, braking systems, parking brakes, and throttle control.

Highly automated "level-two" vehicles will add a suite of capabilities, such as redundant computer backup to ensure network connectivity, basic self-driving capability, and adaptive cruise control that could slow down the vehicle when its intersection assistance program detects merging traffic lanes.

Several companies are experimenting with using level-two automation. These include the fleet of Google Inc. cars navigating the streets of Mountain View, Calif., and the "Freightliner Inspiration Truck," a self-steering 18-wheeler prototype unveiled in May by Daimler Trucks North America LLC. Caravans of these level-two trucks could be cruising U.S. roads as soon as 2020, using vehicle-to-vehicle communications and radar-based active braking systems to follow one another very closely.

To ensure safe operations, the entire chain of trucks could stop very quickly because they would all apply their brakes automatically the instant the lead vehicle detected any sort of collision or obstacle—slowing down more quickly than human perception and reaction time allow, Frost & Sullivan says.

The final step will be "level-three" automation, relying on a fully autonomous vehicle that is capable of performing all trucking functions without a human on board at any point.

"For level three to work, we need not only the vehicle technology but also the communication networks to talk to each other, with a convergence of vehicle technologies like assisted driving with vehicle-to-vehicle communication infrastructure," Vidyasekar said in the webcast.

This full autonomy could be achieved in 10-15 years if current trends continue. Those trends include sophisticated vehicle technologies, the communication infrastructure, the availability of geographic information systems (GIS), and wireless connectivity.

"If there is a clear path for these technologies to develop and for these driving assistance systems to merge with these infrastructure systems, I think automated trucks and automated fleet-management systems could very well be a reality," Vidyasekar said.