When Gen. Mike Murray, commander of the U.S. Army Futures Command, stated in September that Project Convergence is—other than current combat operations—the “most important thing happening in the Army right now,” both then-Secretary of the Army Ryan McCarthy and Army Chief of Staff Gen. James McConville nodded in agreement. The initiative is not only changing the way the Army acquires new technology; it is changing the Army itself.

Born of a brainstorming session by the Next-Generation Combat Vehicle Cross-Functional Team, Project Convergence is an ongoing campaign of learning that brings together new and developing technologies from across the Army enterprise.

Project Convergence centers on delivering data and cloud technologies to the tactical command and is rooted in an overarching requirement to reduce the time needed in combat decision cycles. Project Convergence is aligned with the Army’s goal of becoming fully compatible with the Multi-Domain Operations concept. It is nested within the joint concept, Joint All-Domain Operations.

The project is not about making an incremental change. It is an effort to transform the way the Army operates in all domains—air, land, sea, space and cyberspace—to ensure overmatch against any adversary in future conflicts.

While Project Convergence is obviously focused on technology, it cannot be fully realized without an understanding that people and organizations have contributed to its success.

Harsh Environment

The culmination of Project Convergence for 2020, known as Project Convergence 20, was a six-week experiment with multiple developing technologies at Yuma Proving Ground, Arizona. Yuma was selected for its harsh environment, which challenged both personnel and nascent technologies. The final event of Project Convergence 20 was demonstrations of the experiments held for Army, joint and allied senior leaders on Sept. 21 and 23.

In January 2019, Brig. Gen. Ross Coffman, director of the then-newly created Next-Generation Combat Vehicle Cross-Functional Team, gathered his staff and outlined his vision for future combat vehicles. Specifically, Coffman envisioned commanders, whether in tanks or command posts, having immediate access to all fires and other effects at the push of a button—and that targets should automatically be put on a Common Operating Picture a commander could use to determine what action to take against what target.

From the beginning, it was clear that the use of artificial intelligence and machine learning were going to be the keys to augmenting these largely manual processes from the past. The cross-functional team’s operations officer, Lt. Col. Robert “Jay” Wisham, and myself as the team’s strategic plans officer, were assigned to take the lead and begin to develop what was initially dubbed Project Quarterback.

Wisham and I broke the task into four problems, which eventually became Quarterback’s four lines of effort: 

1. Artificial intelligence-enabled aided target detection/recognition that geolocates and identifies the enemy.
2. Tactical behaviors inherent within their autonomy/artificial intelligence systems that enable them to operate effectively in the environment in collaboration with manned systems.
3. Artificial intelligence decision support agent that analyzes the data and cues sensors and effectors, accounts for the operational plan and guidance, and offers options for execution.
4. Synchronized data management strategy that enables access to the appropriate data, in the proper format, to enable the other three subproblems.

These lines of effort coalesced by May 2019, and the team moved toward finding solutions. An artificial intelligence symposium to tackle the challenge was organized, and the event was advertised across DoD. Any individual or entity anywhere in DoD that had any capabilities that might align with one or more of the four lines of effort was invited. Intentionally, little to no vetting was done of those organizations wishing to send a representative. It didn’t matter how close the tie was, but if the offeror thought it was close, the organization was welcomed to take a seat at the table.

For a week in July 2019 at Fort Belvoir, Virginia, each attending organization presented its relevant projects to the entire audience. Technologies and programs were organized and prioritized based on relevance to Quarterback’s four objectives. The project officers found many sister organizations were developing similar or the same capability. Or, in other cases, no organization was really working on a need, creating a potential gap in Army capabilities.

Line of Effort Leads Chosen

Closing the symposium, four line of effort leads were selected. The leads were individuals whose charter was to look objectively across their line of effort areas to find ways to integrate and synchronize related efforts, regardless of the project’s home organization, and to evaluate commercial or contracted organizations with similar capabilities, bringing them onto the team if warranted.

The leads’ role was to crosscut all science and technology portfolios—regardless of project home organization or DoD branch—to find the best-in-breed of any capability within their lane. A series of “Technology Rodeos” was then held to experiment with and compare similar capabilities, regardless of where they came from, so true, best performers could be found. From there, related technologies began to be integrated into a single system.

The efforts were invigorated by a firm timeline: to hold a large-scale experiment and demonstration within a year to benchmark where the Army stands, compared to a goal of being able to use two sensors to identify two targets and task two shooters to destroy those targets simultaneously. Two sensors, two targets, two shooters. Simple objectives to achieve, coupled with a willingness to risk failure in order to allow maximum learning—then stretch from there.

As Quarterback began to mature, more organizations added complementary capabilities to the effort. The Future Vertical Lift Cross-Functional Team added developing aviation concepts. The Long-Range Precision Fires Cross-Functional Team added the artillery component, and the Assured Positioning, Navigation and Timing Cross-Functional Team brought in key sensor-to-shooter assets. Ultimately, Quarterback became Project Convergence and the premium technology demonstration effort for Futures Command—and for the Army as a whole—that it is today.

Talented Team

Project Convergence 20 was successful in large part because of the talented individuals from across the Army enterprise who attacked the problem. The team included soldiers and Army civilians from the full spectrum of disciplines, from seasoned master gunners to scientists to engineers. The team quickly came together and meshed in Yuma.

Even with the right team in place, however, the event would not have been successful without adherence to several foundational tenets:

• The Project Convergence 20 mantra was to “experiment early and often” in order to synchronize and inform all efforts within the project. Experimentation is a significant portion of the campaign of learning, and major experiments will be held annually to culminate the year’s efforts and inform the campaign of learning. Major demonstrations and experiments will be held in the fourth quarter of each fiscal year and will include capabilities from across the Army.
• Experiments will incorporate operational scenarios against near-peer adversaries in contested environments to ensure Project Convergence remains relevant to the future operational environment and to convey the tactical relevance of the project to future senior decision-makers.
• All potential performers will be brought together through Rodeo events, allowing for qualified and quantified comparisons to be made between each potential solution to a problem. All potential solutions are treated equally, whether they’re Army research laboratories, science and technology centers, other DoD entities, traditional original equipment manufacturers, universities or nontraditional contractors.
• Project Convergence must continue to intentionally consider multiple potential solutions, especially in immature areas, as risk mitigation and to ensure innovative “best performer” solutions can be developed. A broad field of contractor and government solutions may be culled through an initial Rodeo to a shorter list of providers and technologies showing the most potential for continued development, and later culled to one or two specific solutions through a down-select process.
• Knowledge points offer Army senior leaders programmatic control of the information generated within the project through the development and approval of the following year’s learning objectives. These knowledge points provide the opportunity to review the current year’s objectives. Going forward, knowledge points will also offer a culmination of the knowledge generated to date for dissemination to stakeholders of the program.
• Knowledge gained will be cataloged and curated to inform decisions, provide information collaboration across DoD and document the history of the program. The transition of knowledge from the science and technology community to the acquisition community will occur through these knowledge points.

Focusing on 2 Systems

While numerous systems and enterprises were employed during Project Convergence 20, the operation of two systems deserves special attention. The first was the data network that provided the backbone through which all other systems were connected. The second was the FIRESTORM artificial intelligence decision agent. FIRESTORM stands for FIRES Synchronization to Optimize Responses in Multi-Domain Operations.

The network at Project Convergence 20, managed by the Network Cross-Functional Team, took in data from all sensors and quickly moved it to all decision-makers. The event showed that while the network must continue to be made more robust, it can provide the critical link between sensors in all domains, artificial intelligence, commanders and, ultimately, the shooters in all domains.

FIRESTORM is the brain at the center of all systems. It is an artificial intelligence agent that takes the cognitive load off the humans who are ultimately making the fire-no-fire decision. FIRESTORM can process in tenths of seconds what a human might otherwise require tens of minutes to process. To be sure, humans always make the ultimate decision to fire, but they do so armed with recommendations that can be made near-instantly by FIRESTORM.

While this technology remains in its relative infancy, it shows clear promise for the command-and-control needs of the joint force in the near future. The system is being developed under the guidance of the U.S. Army Combat Capabilities Development Command Armaments Center at Picatinny Arsenal, New Jersey.

Mistakes Acceptable

Nearly 1,000 people participated in some aspect of the Project Convergence experiment at Yuma Proving Ground from Aug. 10 through Sept. 25. It should be noted that not every round landed on target, not every aircraft launched on time and not every sensor communicated data back to the network. Those were not the goals of the experiment. The goal was to learn; making mistakes was acceptable as long as lessons were captured from them.

The event showed Army leaders where gaps still exist. The Army learned which systems need to be more robust and which systems are nearly ready for fielding. Up and down the line came the reminder that none of this technology matters if it cannot be put into the hands of soldiers (and Marines, airmen and sailors) in a timely fashion. It was incredibly empowering to see an idea formed from a small-team brainstorming session gain traction so quickly and with such vigor that it ends up causing rippling changes throughout the Army enterprise in less than 18 months.

While the final demonstration of Project Convergence 20 was a monumental event, it too now becomes a waypoint as the Army continues on a path of modernization to ensure it continues to be the world’s preeminent, most technologically advanced and lethal fighting force.

* * *

Matthew Horning is director of Tactical Artificial Intelligence Applications assigned to the U.S. Army Futures Command’s Next-Generation Combat Vehicle Cross-Functional Team, Detroit Arsenal. He served as director of operations for Project Convergence 20. He serves as a lieutenant colonel in the U.S. Army Reserve.