From Unmanned Systems magazine: Military objectives could advance driving systems to full autonomy
Semi-autonomous vehicles traveling in convoys, or platoons, have the potential to save lives and transport material more efficiently and safely than manned vehicles, but the technology has a way to go before it can be touted as fully autonomous.
In the commercial sector, the rush to full autonomy is limited by fuel cost efficiencies already achievable with semi-autonomous systems, but the need to protect soldiers on the ground and in war zones may drive the U.S. military to advance the technology into full automation sooner rather than later.
Technology developers and the U.S. military are testing the limitations of semi-autonomous transport vehicles traveling in succession, known as platooning, in conditions similar to those in which the military is currently engaged in conflict. In the military, the partially unmanned technology has the potential to reduce casualties by reducing the number of personnel needed to transport materials from one area to another, limiting human exposure to enemy attacks.
With the potential for full automation to give the U.S. the upper hand by preventing the loss of soldier lives, military leadership sees an incentive to further develop the technology. Minimizing the number of drivers means figuring out how the systems work with only one driver in the lead vehicle and no driver in subsequent vehicles until full automation is more tangible.
In testing, the one-driver, semi-autonomous platoon has encountered some problems — detect and avoid sensors and software come with various strengths and weaknesses and, depending on the climate and conditions of the military activity, may not always be as effective as human observation and judgment, says RAND Corp. research Shawn McKay. That’s why RAND is recommending continued research on the minimally manned approach, in which “we can keep one soldier in the truck and collect more data to improve the system so we can increase the capability over time.”
McKay points to the research that Google’s self-driving system Waymo is conducting in California, in which it is collecting data on semi-autonomous vehicle systems. He says Waymo is specifically observing when the driver feels the need to take over the vehicle, and why. These human behavioral observations are critical, particularly in a war zone, to understanding how the technology can better identify and adapt to not only the human operator but to real-time conditions on the road or route.
Lockheed Martin is one of the contractors working with the U.S. Army to test its own autonomous driving kit on military transport vehicles. Lockheed’s Autonomous Mobility Applique System (AMAS) kit provides a retrofit for existing vehicles, allowing for semi-autonomous capability in leader-follower convoy operations. In the “driver warning/driver assist mode,” AMAS assists a driver in tasks such as avoiding obstacles and collisions and maintaining a safe distance from the vehicle ahead and a steady position in a driving lane. And “leader-follower mode” allows personnel to link a large number of vehicles together in a convoy so the follower vehicles can operate without a person in the driver’s seat.
AMAS logged more than 55,000 testing miles during the U.S. Army Extended Warfighter Experiment at Fort Leonard Wood, Missouri, and Fort Bliss, Texas, last year “in a variety of mission scenarios," says Lockheed Martin Combat Maneuver Systems Director Kathryn Hasse. Testing of the AMAS system during the warfighting experiment involved the use of a “palletized loading system” vehicle convoys in which the lead vehicle was driven by a soldier and three to four vehicles without drivers followed robotically.
“Soldiers operating the AMAS vehicles provided us very positive feedback about how the system freed them up to do the job of a soldier instead of the job of a truck driver," Hasse says.
But the kits are equipped with heavy sensors, cameras and other upgrades that put some limitation on aging military vehicles. U.S. Army Chief of Staff Gen. Mark Milley told the Senate Armed Services Committee on April 12 that the vehicles are “maxed out” and there’s little room available to make any additional upgrades. While for the military that might mean updating the vehicles themselves, for technology developers there’s still a lot of prototype testing that will need to be done not just to optimize detect and avoid technology, but to do so while lightening the load to enable a vehicle to operate autonomously, efficiently and safely.
Milley told lawmakers that the military branch is embracing a next-generation fleet of semi-autonomous vehicles.
“Every ground and rotary wing vehicle that the United States Army produces” will be either fully autonomous or semi-autonomous, he said. “We’re in a period of changing character of war and [artificial intelligence and robotics], perhaps more than other technologies, will have a fundamental impact,” Milley testified.
Milley added that the service branch is also working on different variants of the technology including for tanks and medical transportation. Through the modernization effort, the Army intends to “eventually replace the entire family of vehicles that we have.” So it’s no surprise Lockheed hopes to steer the AMAS program into new production logistics trucks.
Milley also said the Army is in the process of developing experimental prototypes and expects to deploy the systems by 2028.
The AMAS technology is also being propagated into the U.S. Army’s Autonomous Ground Resupply Program, which Lockheed anticipates will extend AMAS into fully autonomous operation. AMAS is also the foundation for other autonomous vehicle products in development by Lockheed Martin for industrial and commercial operations.
A convoy of autonomous military vehicles using Lockheed’s AMAS system take part in a demonstration. Photo: Lockheed Martin
Trucking sweet spot
In the commercial space and across U.S. roadways, semi-driverless convoys can help reduce man hours and accidents and more efficiently transport goods across state lines, but in this sector the fuel efficiency benefits are substantial enough to dampen the transport industry’s demand for full automation.
Most of the commercial technology under development is still in the level one phase, meaning a driver is still present in each vehicle. Current testing in platooning of semi-autonomous trucks involves one driver in the lead vehicle and one driver in the follower vehicle.
Technology developers Freightliner and Peloton Technology are expected to take their semi-autonomous driving systems to market this year. The systems include driver-assisted automation equipped with collision avoidance software and vehicle-to-vehicle communication and synchronization to enable the second driverless truck to travel in close succession to the first manned, semi-autonomous truck, allowing for greater realized fuel economy savings.
But there are fundamental differences between the two technologies. Peloton offers a kit that can be equipped in any transport vehicle. Peloton’s vehicle-to-vehicle (V2V) communication system uses the cloud to track V2V-equipped trucks and enable them to communicate with one another and begin platooning at any point that two vehicles find each other. Freightliner, on the other hand, offers a fully integrated Freightliner truck equipped with platooning capabilities, but only with other Freightliner trucks.
Peloton uses the V2V communication to connect the braking and acceleration between the two trucks, allowing the lead truck to control the acceleration and braking of both trucks simultaneously. This allows the vehicles to travel in close proximity, which would otherwise be extremely dangerous, and reduces the aerodynamic drag for both the trailing and the leading truck. Independent fuel efficiency testing by the U.S. departments of energy and transportation as well as the North American Council for Freight Efficiency (NACFE) validate the claim.
“Without question, truck platooning is a valid method of reducing fuel consumption for tractor-trailers engaged in long-haul applications,” according to a 2016 report from NACFE. “Once the trucks have moved into close following distances, all of the engaged vehicles receive a significant fuel economy boost thanks to increased aerodynamic efficiencies. The lead vehicle, which bears the brunt of the aerodynamic load, typically sees only a modest fuel economy boost. But the trailing truck in a platoon, which is now operating in a low air pressure aerodynamic ‘sweet spot,’ can see significant increases in fuel economy performance at highway speeds.”
In fact, the majority of the monetary benefit of autonomous technology in truck platooning is realized in automating the brakes and throttle, according to industry consultant Richard Bishop of Bishop Consulting.
“You don't get any more, or monetary, benefit if you automated steering, for instance, because it's all about these trucks moving close to each other, following each other closely,” Bishop says. “It's just sort of a low hanging fruit.”
For Freightliner, that low-hanging fruit focuses on creating the safest and most efficient experience for the driver. Freightliner’s highway patrol software processes information from camera and radar technology to enable lane stability, collision avoidance, speed control, braking and steering. The information is distilled and displayed on a dash interface to communicate the truck’s performance to the driver, while also accepting commands. And video displays replace exterior mirrors, reducing blind spots.
While semi-autonomy may be optimal in the commercial transportation industry for now, there are other advancements technology developers are exploring, including adding more follower vehicles to the convoy and eliminating the need for more than the one driver in the lead vehicle.
Semi-autonomous platooning is currently legal in at least 13 states, but Bishop emphasizes more are on the way and the industry does not see regulation as a huge hurdle to semi-autonomous commercialization of platooning trucks.
Below: A measurement of how platooning can save fuel and money. Image: Peloton Technology