9 septembre 2019 | International, Aérospatial
Jumping into algorithmic warfare: US Army aviation tightens kill chain with networked architecture
By: Jen Judson NAVAL AIR WEAPONS STATION CHINA LAKE, Calif. — In the skies above China Lake, California, from the back of an MH-47 Chinook cargo helicopter, an operator with a tablet takes control of a Gray Eagle drone and tasks it with firing a small, precision-glide munition at an enemy target located on the ground. But at the last second, a higher level threat is detected and the munition is rapidly redirected toward a different threat, eliminating it within seconds. This was made possible through the architecture, automation, autonomy and interfaces capability, or A3I, built by the Army's Future Vertical Lift Cross-Functional Team under Army Futures Command. The demonstration showed the ability to nimbly pass control between operators of unmanned systems and munitions through a networked architecture of systems also receiving and filtering real-time, pertinent information to aid in operational decision-making. “It was our first jump into algorithmic warfare,” Brig. Gen. Wally Rugen, who is in charge of the Army's FVL modernization effort, told Defense News following the demonstration. “We definitely didn't jump into the deep end of the pool, but we jumped in and, again, we are into pursuing that as far as we can take it to help soldiers be lethal.” The Aug. 26 demonstration sought to tighten the kill chain and allow for more advanced teaming between air assets and troops on the ground using a resilient network. “When you talk about our kill chain, we are trying to take seconds out of our kill chain,” Rugen said. “We feel like we understand the reverse kill chain — the enemy coming to get us. Our kill chain is going to get them, and we want our decision-making to be as precise and as expeditious as possible,” using automation and autonomy, he added. AI3 was developed over the course of nine months and culminated in the demonstration at China Lake. "Going from a concept, and in a matter of months putting it into an experiment: That was probably the most impressive thing, particularly if you look back at the history of how we do these,” James McPherson, the official performing the duties of the undersecretary of the Army, told Defense News. McPherson attended the demonstration to emphasize the importance to senior Army leadership of modernization efforts within the service. The FVL effort in particular includes ensuring manned, unmanned, munition and other air-launched effects are all seamlessly networked together to fight in advanced formations in a congested environment, such as an urban area, and that they are prepared to fight across multiple domains. Using an interface called Arbitrator, the service networked together a variety of targeting identification and rapid automated processing, exploitation and distribution, or PED, capabilities as well as real-time weather information and several other features and capabilities to help operators of unmanned systems penetrate, in the case of the demonstration, an urban environment. AI3 in action During the demo, one of the systems integrated into the network tied to a ground sensor detected a possible threat on the ground. Seeing the threat detected in the system, a helicopter pilot then gained control of an extended-range Gray Eagle and tasked it to perform reconnaissance of the possible target. Using the UAS, the pilot identified the threat as an enemy surface-to-air missile system. The pilot then ordered the UAS to fire a Dynetics GBU-69 small glide munition to defeat the target, marking the first time the munition had been fired from a Gray Eagle. But as the munition closed in on the target, the system picks up on another threat deemed more important for elimination. The information for this decision came from the integrated PED systems that use machine-learning algorithms to accurately identify items of interest. Another operator then redirected the munition during its final seconds of flight to hit the new, more pressing threat. Why does the Army need A31 capability? To build the system, the government took the lead integration role, Chief Warrant Officer 5 Cory Anderson, the UAS branch chief for Army Special Operations Aviation Command, said at the demonstration. This ensured the service's ability to get the right levels of interoperability between subsystems. But almost all of the capabilities tied into the government's black box came from small businesses and academia. Much of the initial development has come from the special operations side of the house. The demonstration was viewed from a tactical operations center, with screens lining the walls of a large air-conditioned trailer, but the system has a scalable control interface and can be remotely accessed from a cockpit or even a tablet used by a soldier on the ground. This breaks the Army free from having to use a ground control station, Anderson said, meaning the footprint and logistics tail can be drastically reduced. To put together the tactical operations center and ground control station, it took roughly seven C-17 planes to move heavy equipment into China Lake. “We can't sustain that,” Anderson said. “We believe we can get it down to a two C-17 load-out just by minimizing the generational requirements alone.” By integrating PED systems that use machine learning into A3I, the Army no longer requires a large number of people — roughly 30 at a time — to conduct PED from full-motion video. The Arbitrator system allows for operators to pass control of various systems back and forth at different levels of control, from just receiving information from a sensor or UAS to controlling a payload to the entire system. The system is also under development to improve its automation levels. The utility of passing control to a relevant operator not tied to a ground station means taking out the middle man that doesn't have the same advantageous access to the tactical edge another possible operator might have. Rugen said that if there's an operator on the ground close to the action, it's much easier to take control of systems rather than try to direct someone far away to the right location to get eyes on a possible point of interest or target in order to make an actionable decision. “What if the squad leader could just grab the sensor because we have the hierarchy?” Rugen noted. While the capability was developed and demonstrated by the FVL Cross-Functional Team, the system has applications for almost everything on the battlefield, from applications to long-range precision fires targeting capabilities to next-generation combat vehicle teaming to soldier systems. Both directors for the Long-Range Precision Fires and the Network cross-functional teams were present at the demonstration. While the unclassified version of the demo didn't show capability, the classified version addresses the architecture's capability to protect itself against threat-representative electronic attack. “We want to make sure we have a resilient network,” Rugen said. The next step is to move the Arbitrator system onto an airborne platform, which would completely eliminate the ground control station. That will be demonstrated in roughly a year. https://www.defensenews.com/land/2019/09/05/jumping-into-algorithmic-warfare-army-aviation-tightens-kill-chain-with-networked-architecture/