Army tests robot-force attack on enemy 'tank ditch, minefield'

Surging into heavy enemy fire, navigating rigorous terrain and forward-placing sensors and weapons close to enemy targets, teams of Army robots conducted a “deep assault through a breach” during an exercise intended to prepare the service for a new kind of man-machine drone warfare.

The Army exercise, which pitted groups of unmanned vehicles or ground drones against a mock enemy “tank ditch” and “minefield,” was part of a massive service-wide modernization effort to prepare for a new generation of combat -- one wherein self-navigating drones directly confront enemy fire in high-threat war scenarios while humans perform command and control at safer distances.

During the Army demonstration, which took place several months ago, there … “was not a single soldier in any vehicle” conducting the initial breach, Commander of Army Futures Command, Gen. John Murray, told reporters.

Murray described the exercise at Yakima Training Center, Washington, during which soldiers in manned vehicles and control centers conducted an attack at safer distances, as an extremely useful opportunity to learn about how autonomous systems are changing warfare -- and talk to soldiers about the experience.

“In my mind what is critically important is what does this formation do in terms of how soldiers operate differently and what it does with man-machine interface. What is the cognitive load? Can soldiers control four or six autonomous vehicles?” Murray said.


The exercise, designed to closely replicate the circumstances of a high-risk assault, included 10 unmanned vehicles and 6 manned vehicles, service officials said.

“The obstacles were complex objects that included wire, tank berm and a tank ditch. There was an OPFOR (opposing force) which released smoke … but the event was solely tied to breaching the complex obstacles,” John Jolokai, Deputy, G-3, NGCV Cross Functional Team, Army Futures Command, told Warrior in a written statement.

The developmental concept, as articulated by Murray, is to ease the “cognitive burden” for human commanders operating in a command and control capacity at safer ranges less vulnerable to enemy fire; sensor data, targeting, forward operating electronic warfare or even certain kinds of cyber-controlled fire controlled systems, can be gathered and organized for human decision-makers - who can then expend crucial energy upon dynamic, pressing warfare decisions. While state of the art autonomy is rapidly improving its ability to perform critical procedural functions, the unique attributes of human cognition and decision-making are still needed for problem-solving, adjusting to fast-changing variables and, of course, making decisions regarding the use of lethal force in warfare.

Strategically and tactically speaking, autonomous or robot warfare brings a host of advantages above and beyond allowing soldiers to operate at safer stand-off distances. Computer processing speeds, fortified by AI, are increasingly able to assimilate and organize vast amount of combat-relevant information. For instance, emerging algorithms are able to monitor hours or even days of surveillance footage and identify key moments of relevance without requiring otherwise time-consuming human efforts. This can be crucial, not only because it frees up human energy for more important tasks, but because it brings massively increased speed and data organization by virtue of using advanced computer programs. AI, and machine learning, work by assimilating new information, comparing it against what seems like a limitless existing database, and performing millions of analytical functions in seconds.


Advanced AI is able to perform some kinds of analytics in near-real-time, wirelessly, to give human commanders lifesaving information amid the fog of war. This kind of AI-enabled autonomy can not only solve problems but discern crucial targeting or sensor information, fuse various streams of input together or even expedite networking between land, air and sea assets operating together in joint warfare. Along these lines, Murray cited upcoming multi-domain wargames involving Army, Navy and Air Force senior weapons developers intended to further develop this kind of approach.

Also of key significance, AI-oriented autonomous platforms can greatly shorten sensor-to-shooter time and enable war commanders to quickly respond to, and attack, fast emerging moving targets or incoming enemy fire. Given all of these variables, coordinated autonomous air and ground drones are naturally, according to Army weapons developers, indispensable to the demands of future war.

“Tele-operation of robots is a baseline of current technology. That is sufficient for some operations but not sufficient in terms of where we want to go,” Dr. Bruce Jette, Assistant Secretary of the Army - Acquisition, Logistics and Technology, told reporters.

Many kinds of tele-operated robots have been used in combat for quite some time, such as Iraq and Afghanistan; the PackBot, for instance, is a small, human-operated sensor designed to surveil high-threat forward locations and, in many instances, use electro-optical or seismic sensors to detect IEDs. Some of the technology used to operate these have advanced to what is called a kind of “semi-autonomy” wherein the robot is still controlled by a human yet able to perform certain key functions without needing human intervention. This, Army and industry developers explain, was done through various software upgrades. PackBots and other unmanned systems, which at times were deployed by the thousands in theater, are credited with saving many lives in combat.

Various kinds of advanced autonomy, naturally, already exists, such as self-guiding aerial drones and the Navy’s emerging “ghost fleet” of coordinated unmanned surface vessels operating in tandem. Most kinds of air and sea autonomous vehicles confront fewer operational challenges when compared to ground autonomy. Ground warfare is of course known to incorporate many fast-changing variables, terrain and maneuvering enemy forces - at times to a greater degree than air and sea conditions - fostering a need for even more advanced algorithms in some cases. Nevertheless, the concepts and developmental trajectory between air, land and ground autonomy have distinct similarities; they are engineered to operate as part of a coordinated group of platforms able to share sensor information, gather targeting data and forward-position weapons -- all while remaining networked with human decision-makers.


The Army expects manned-unmanned teaming and autonomous drone combat to figure prominently in its future armored combat vehicle plans; its now-in-development next-gen infantry fighting vehicle is being referred to as an "optionally manned" vehicle, and senior weapons developers say most, if not all, future combat vehicles will be engineered to operate with nearby ground and air drones.

Murray’s explanation of machine interface is reinforced by an interesting 2015 essay in the “International Journal of Advanced Research in Artificial Intelligence,” which points to networking, command and control and an ability to integrate with existing technologies as key to drone-human warfare.

“They (drones) should effectively interact with manned components of the systems and operate within existing command and control infrastructures, to be integral parts of the system,” in Military Robotics: Latest Trends and Spatial Grasp Solutions, by Peter Simon Sapaty - Institute of Mathematical Machines and Systems, National Academy of Sciences.

Increased use of networked drone warfare not only lowers risks to soldiers but also brings the decided advantage of being able to operate in more of a dis-aggregated, or less condensed formation, with each drone and soldier system operating as a “node” in a larger integrated network. Dispersed forces can not only enable longer-range connectivity and improved attack options but also reduce force vulnerability to enemy fire by virtue of being less aggregated.

“Despite the diversity of sizes, shapes, and orientations, they (drones and humans) should all be capable of operating in distributed, often large, physical spaces, thus falling into the category of distributed systems,” Sapaty writes in the essay.


Also of great significance, Army thinkers explain, is that greater integration of drone attack assets can streamline a mission, thereby lessening the amount of soldiers needed for certain high-risk operations.

A paper in an Army University Press publication explains how drones can “expand” the battlefield. By utilizing drone systems for combat… “fewer warfighters are needed for a given mission, and the efficacy of each warfighter is greater. Next, advocates credit autonomous weapons systems with expanding the battlefield, allowing combat to reach into areas that were previously inaccessible,” the essay states. (Amitai Etzioni, Ph.D., Oren Etzioni, Ph.D.)

Murray stressed that the impact of autonomous systems upon the requirements for future threats is so substantial that it needs to not only remain within the realm of concept but also inform future Army doctrine.

The Army acquisition community and Army Futures Command are working together to accelerate these kinds of technological advances through close collaboration with industry and academic developers. Futures Command, for instance, is now working with both the University of Texas and Texas A&M from its Austin, Texas headquarters.

Army officials say a number of upcoming experiments are planned for the near term as precursors to a drone-soldier attack-breach demonstration slated for 2023.

The next goal, to be attempted in an upcoming simulation is -- according to Murray --- the following:


“Show me how to conduct a breach of a complex obstacle without a single soldier within a kilometer or two of the breach site,” Murray said.