WASHINGTON — The U.S. Army Research Lab made breakthroughs this summer on two neural networks projects that could assist commanders’ decision-making on the battlefield and provide soldiers’ health information through fibers in their uniform.
The advancements come as the U.S. military is preparing for data-driven battle, in which gobs of data are transmitted across the battlespace, processed and used in a commander’s decision-making. Neural networks are a combination of algorithms that work together to recognize patterns in data through a process similar to that of the human brain.
The first project is working to provide a tool to battle commanders that quantifies uncertainty in data analysis using neural networks. Researchers associated with the Army Research Lab created a new framework for neural network processing that would use artificial intelligence to provide confidence ratings. The tool could enhance the ethical use of AI and provide commanders with a new level of confidence in their decision-making.
“The idea is to provide some sort of uncertainty measurements from the AI system so that the human counterparts are able to acknowledge and realize that the AI is not going to be perfect, but we can understand sort of how certain it is, and then use that information to make more informed decisions,” said Maggie Wigness, Army researcher and deputy collaborative alliance manager of the Internet of Battlefield Things Collaborative Research Alliance (IoBT CRA).
Several factors influence uncertainty around battlefield data and AI outputs, Wigness said. Data could be manipulated, it could have random “noise” that messes with the algorithms or uncertainty could stem from a degraded sensor that an adversary attacked. To quantify uncertainty in the future, researchers categorized sources of uncertainty in military networks, reviewed frameworks to represent the unknown and created solutions to manage it.
As the Army works to enable multi-domain operations, in which the service contests adversaries across the five warfighting domains, data from across all five will have to be fused and integrated. The neural network work fits into the IoBT alliance’s work that is “focused on processing data and establishing connections and communicating across a large network of sensors, those sensors could be sort of anywhere across these multi domains that need to then be fused together,” Wigness said.
Meanwhile, Army researchers at the service’s Institute for Soldier Nanotechnologies at the Massachusetts Institute of Technology developed a programmable fiber that could be sewn into soldier’s uniforms in the future and provide biometric data that would help monitor soldier health. That fiber, according to the service, can sense, store, analyze and infer activity. The digital fiber technology is in the proof of concept phase, demonstrating that the fiber has memory capabilities. Researchers loaded a movie onto the fabric to show that it could store mass amounts of information.
In the future, researchers want the fiber to be able to store algorithms that can analyze the data that it collects. According to an ARL post about the breakthrough, the fiber has a neural network of 1,650 connections and successfully collected 270 minutes of surface body temperature data. Using artificial intelligence in the lab, the fiber identified what type of activity the soldier was undertaking with 96 percent accuracy.
Eventually, researchers want the fiber to be able to generate and store power to run artificial intelligence analysis or communicate with sensors and communications systems.
“You can imagine if a soldier is non-responsive, we may be able to have communication with these wearable computers ... where the soldier’s uniform is communicating with Central Command or their home base on the battlefield,” said James Burgess, Institute for Soldier Nanotechnologies program manager for the research lab. “So if they’re nonresponsive, we may know ... whether they’re bleeding, whether they’re breathing, what their pulse rate is, so we would have some diagnostic information before we even reach the soldier with the rescue team.”
The capability is still “very early stage,” Burgess said, and the team is aiming to have an impact on the battlefield by 2050. The capability will be an important piece of providing soldiers information on their physical state as well.
“With the adrenaline flowing, it’s really important for soldiers to know what’s happening to them, their physiological state, and have warning buzzers, essentially, that go off [if] they’re overheating or dehydrated, if they’ve been exposed to something that is beginning to impair their cognitive or physical abilities,” Burgess said.
Andrew Eversden covers all things defense technology for C4ISRNET. He previously reported on federal IT and cybersecurity for Federal Times and Fifth Domain, and worked as a congressional reporting fellow for the Texas Tribune. He was also a Washington intern for the Durango Herald. Andrew is a graduate of American University.