AI needs of U.S. Army Futures Command over the next five years

The U.S. Army expects it will need a range of AI-enabled systems to operate on the battlefield of the future. (T’Jae Ellis/US Army Illustration) (Army Research Laboratory)

WASHINGTON: The U.S. Army Futures Command has outlined 11 areas of artificial intelligence research it is interested in over the next five years, focusing on data analytics, autonomous systems, security and decision assistance.

The U.S. Army and Department of Defense are working to connect sensors and shooters on the battlefield, as the U.S. Army Futures Command, headquartered in Austin, Texas, made a wide-ranging agency announcement. By analyzing data and assisting commanders in the decision-making process, artificial intelligence will be key to this effort.

The U.S. Army is “particularly” interested in artificial intelligence research on autonomous ground and air platforms, “which must operate in open, urban and cluttered environments,” according to a bulletin issued by the command’s Artificial Intelligence Integration Center. The bulletin document specifically Requires research into technologies that allow robots or autonomous systems to move in urban, confrontational environments, and technologies that reduce the electromagnetic distribution of the system. It also wants to learn more about artificial intelligence that can perceive fuzzy targets and understand terrain obstacles.

The announcement document identifies several needs related to data analytics over the next five years. The U.S. Army is interested in human-machine interface research, and further research is needed to predict adversary intent and behavior on the battlefield. In the same category, the U.S. Army wants to be able to fuse data from disparate sources and have analytical capabilities to “leverage” classified and unclassified sources to create “enhanced” intelligence products.

The Army also wants to be able to combine human insight with machine analysis and develop improved ways to “effectively” communicate analysis results to humans.

“The U.S. Army is interested in research in areas of artificial intelligence/machine learning that can reduce the cognitive load on humans and improve overall performance through human-robot collaboration,” reads the announcement document.

Likewise, the U.S. Army needs to do more research over the next five years on how to better Display data to humans. Data must be clearly presented to users—for example, through charts or graphs—so they can understand the meaning of the information.

“The U.S. Army is interested in research that can improve situational awareness and visualization and navigation of large data sets to enhance operational activities, training, and readiness,” the bulletin document reads. Along the same lines, the U.S. Army is also pursuing new approaches to visualize sensor data and large datasets with multiple sources.

The U.S. also wants more battlefield perception research into AI, including detecting people, equipment and weapons, even when occluded. It wants to perceive these objects based on “physical, behavioral, cyber or other characteristics”. In addition, the U.S. Army needs artificial intelligence sensors and processors that can detect chemical, biological, radiological, nuclear and explosive threats.

Network and communications security is another area where the U.S. Army wants to do more research. The United States is seeking more research into autonomous cyber defenses and AI-based approaches to offensive cyber capabilities. It also requires novel cyber protection technologies and approaches.

Additionally, to prepare for a future potential GPS-denied environment, the U.S. Army is interested in researching algorithms and techniques to fuse location, navigation, and time sources to provide a “robust” capability.

The Internet of Things, or the vast network of devices connected to the Internet, presents more AI needs for the U.S. Army. According to the request for comment, the U.S. Army is interested in artificial intelligence research, namely “new approaches to enable secure, resilient, and autonomously managed IoT networks in highly complex, hybrid cooperative/adversarial, information-centric environments.”

“The U.S. Army needs to better integrate a wide range of capabilities and equipment and leverage commercial developments in the industrial and human Internet of Things,” the call for proposals said.

Andrew Eversden: Andrew Eversden covers all of C4ISRNET’s defense techniques. He previously covered federal IT and cybersecurity for The Federal Times and The Fifth Domain, and was 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.