What Worked, What Didn't at Army’s Second Connect-Everything Experiment

YUMA PROVING GROUND, Arizona—If last year’s edition of the U.S. Army’s massive connect-everything experiment was a proof of concept, this year’s was a far bigger effort to see just how much those data-sharing concepts might accelerate major military campaigns.

“I’ve seen exponential progress since last year,” said Army Chief of Staff Gen. James McConville at a media roundtable here on Tuesday. “What I’ve seen is the ability to move data. The ability to have speed, range, convergence to get the speed for decision dominance [was] significantly improved.” 

In just its second year, Project Convergence has become the U.S. military’s most important experimentation effort for testing out new technologies for joint all-domain command and control, or JADC2. This year’s version featured 110 technologies, triple that of last year. It involved more personnel, including 82nd Airborne Division troops and others from Navy, Marines and Space Force and drew in participants from Fort Bragg, North Carolina, to the White Sands Missile range, and yet it simulated an even larger battleground in the INDOPACOM area of responsibility.

This year’s event also drew a bigger audience, with members of the Five Eyes intelligence alliance flying in to witness the demonstration in addition to McConville and Army Secretary Christine Wormuth.

Here’s what Army officials said went well this year and what still needs solving. 

What Worked

AI-enabled reconnaissance 

The military has been experimenting with teams of ground robots and quadcopters for reconnaissance for years. But this year’s Project Convergence experiment saw an important step: much smarter and more independent robots. 

Four autonomous ground robots teamed with two tethered drones to conduct a reconnaissance mission with virtually no human control. 

​​“We’ve advanced to the point now where we say ‘Here’s your area. Go ahead and conduct reconnaissance,’” said ​​Col. Andre Abadie, who led the Project Convergence 21 operational planning team. “It’s kinda like giving intent. And so those robotic vehicles and their ability to collaborate and sense decided how to recon that area on their own. And then decided how to report back those threats” to the communications component. “We hope to take that next year and actually do that at night.”

Project organizers also demonstrated that a fully autonomous UH60 helicopter, flying DARPA’s Aircrew Labor In-Cockpit Automation Software or ALIAS, could launch loitering drones like missiles. 

Faster sharing of threat data across services

“We found some challenges with this idea of how do we as services share information and maintain a single view of what that common operational picture looked like. We’ve worked through that, started to push a little bit,” said Abadie. 

Key to that was getting better participation from the Joint Staff, which sent a team to Yuma to debug messaging formats and look at the bureaucratic, process, and policy barriers that slow interservice data sharing. 

“The Joint Staff sent a team out here to sit down with the operators, sit down with systems and say, ‘Hey. Here are the message formats. These are all joint message series. But when the Navy uses it, they do this. When the Marines use it they do this. When the Army does it they do this. If you were to add these three fields, we could all do this in a coherent way together,’” Abadie said.

The exercise also paved the way for the Navy and Marines to integrate sensor data into the Army’s Integrated Air and Missile Defense Battle Command System, said Gen. John Murray, the head of Army Futures Command. 

Better networking via satellites and drones

This year’s event tested new concepts for high-bandwidth mesh networking where adversaries might deny communications—key to the military's evolving vision of highly-networked weapons, vehicles and soldiers. Data was passed via drones and a small constellation of low-earth orbit satellites

Working with soldiers

While last year’s exercise featured mostly scientists and some battlefield coders, this year’s featured actual operators testing new equipment, such as the IVAS headset in a simulated air assault. The headset allows soldiers to better train on real-world situations in a simulated environment. 

“Soldiers were able to do rehearsals. They were able to visualize the objective they were going to assault, go through rehearsals under the leadership of their platoon. Then, as they boarded the aircraft, maintain situational awareness of that objective,” said Abadie. 

More importantly, bringing in actual operators allowed for important feedback not only on how new technologies were working but how well they worked together, said McConville. 

What Needs Work

Battle-ready autonomy

Autonomous robots that can perform reconnaissance missions are cool. But if they rely on the kind of laser-range finding, or LIDAR, technology that today’s self-driving cars use, they’ll be easy targets. The autonomous drones in this exercise needed both LIDAR and GPS. That represents a vulnerability. 

“I can’t use LIDAR on the battlefield. It lights me up like a Christmas tree. I’ll die soon,” said Abadie. The Army, he said, is “slowly advancing this idea of off road autonomous vehicles...We have to use alternative ways of sensing.”

Defending against electromagnetic countermeasures 

Outside of the Tuesday demonstration, the Army took five days to run through seven scenarios relevant to a war involving China. One of those five days, for each scenario, looked at how well the network performed under attack. Both China and Russia possess sophisticated electronic warfare capabilities.

“Doing those type of experiments out here helped us to understand the tradeoffs involved. If you apply certain techniques to make your mesh wave form more resilient you might take a penalty on the range or the capacity,” said Col. Eric VanDenBosch, the chief of staff of operations for the Army’s Network Crossfunctional Team. An understanding of what those tradeoffs look like will shape future Army exercises and weapons. 

Information sharing and situational awareness

While this year’s demonstration showed that the Army, and the military as a whole, has made progress in interconnecting weapons, vehicles, and systems across services, it still represents a carefully planned demonstration. In fact, integrating highly complex jets, drones, ships, and missiles within just one service is incredibly difficult. 

“It goes back to: Army built theirs, the Air Force built theirs. Navy built theirs. The new technology is really interesting, but the real challenge is: ‘How do you begin to tie billions of dollars worth of investment together,” Murray said.  “So it’s the enduring material that everybody has…how do you begin to bring that together?” 

Wormuth offered a sobering take on the road ahead. “We aren’t going to be able to get away from that, maybe not ever. Think of our own reserve components...They are going to have some enduring capabilities while the active Army has some of the signature modernization capabilities so, even inside of our own formations, we will have to think about retroactive, if you will, the ability to remain interoperable. And we will have to do the same with allies and partners. We will never be able to start with a clean sheet of paper.”

That continued challenge is no joke. It’s part of the reason why the Air Force’s Joint All Domain Command and Control experimentation—the Advanced Battle Management System on-ramp experiments—were essentially defunded by Air Force Secretary Frank Kendall this year. 

It’s a problem that speaks directly to the very concept of joint-all domain command and control, what it is, and what it would actually look like. 

“Each of the services has systems to maintain situational awareness,” Murray said. “Getting them to integrate and talk together and how procedures, policies that different services use, when to display what, that’s a problem. Right now, if we wanted a joint situational awareness picture, you would probably be looking at four different screens… so how do you begin to collapse so that a joint commander has everything he can see across the battlespace, what the joint, situational awareness looks like.”

Further complicating that are the rules surrounding secrecy. One senior Defense official speaking on background said that classification issues were going to be a major focus of the Defense Department in the coming year, precisely because overclassification threatens timely information and data sharing. 

Next year’s event will likely feature additional elements and geographic locations, further raising the stakes not only for the Army but for the military’s entire approach to joint all-domain command and control.