DCS 2019: The art of Mosaic Warfare

17 Apr 2019

Agile approach lets military “think small and very fast,” says DARPA director Tim Grayson at DCS plenary.

From Ford Burkhart in Baltimore

DARPA, the $3.5 billion agency for US military research, is employing a new approach to working with industry. The aim here to develop small-scale, instantly useful military tools, a plenary speaker said Tuesday at SPIE’s Defense and Commercial Sensing meeting in Baltimore, Md, U.S.

The system, known as “mosaic warfare,” will allow the US military’s tools to be distributed, heterogeneously, a so-called “system of systems”, allowing the military to work fast and at a large scale to carry out its operations, said Tim Grayson, director of the Strategic Technology Office at DARPA (Defense Advanced Research Projects Agency).

“We can create a new military capability working at the speed of operational planners, as opposed to traditional acquisitions and technology development, so we can have more effective sustained military capabilities against potential competitors,” he said.

Mosaic warfare – involving this system of systems, or SoS – will let the military build new “minimum viable products” and infrastructure needed to carry out the new kind of warfare. “We are planning these complex architectures for interoperability – for example, to create sensors with data links that will let different systems talk to each other. We are using technology to help technologies talk to each other.”

One challenge is to help people transition to new machines. “We want to make it easier for humans to work with new architectures and capabilities they have never been exposed to before,” Grayson said.

In a traditional DARPA program transition, it would do a technology demonstration, and the commercial world would make it work as a mass scale, “say, tens of thousands of sensors, or radios, or looking for something immediately useful, and “you would not have to bet your whole future on it.”

‘Affordable risk’

With mosaic warfare, a product might be created for one user, for one unit in the military, and treated as a pilot. “But it would be at such a small scale,” Grayson said, “the bet they are making is very small. So you can afford to take that risk.”

One example of this thinking about human-machine interactions is a program called RSPACE, standing for Resilient Synchronized Planning and Assessment for the Contested Environment. It is created software for command and control decision tools, helping individuals make a transition to various apps. Individual tools can then go to small groups of users, at the unit level.

Another approach in the SoS tradition is called STITCHES, for SoS Technology Integration Tool Chain for Heterogeneous Electronic Systems. The software allows users to build bridges to translate quickly between various data formats. In turn, the STITCHES system is proving valuable in tests, called Gauntlets, which are a kind of hackathon where, in one week, users, government officials and vendors try to create a whole new software architecture.

DARPA has now perfromed eight such tests, run by Apogee Research in Arlington, Virginia. And at the end, users have taken home the new software, the “glue” from the Gauntlets, and began using it operationally, Grayson told the conference. “They don’t have to build a new system. The just took home the blue software and have a new capability for repurposing things they already have in their arsenal. They are like geek squads, working on the fly.”

‘Leveraging best practice’

To succeed in multi-domain combat, DARPA, Grayson said, is “really trying to leverage the best practices of commercial industry. We are trying to modernize to take advantage of speed.”

In that world, he said, “time to market is more important than making one individual product. It’s about constant adaptation and constant speed. For warfighters who have to go out and use these things, it just dies under the complexity of a top down model,” Grayson said.

Developing communications for mosaic warfare involves not just humans to humans but also sensors going directly to shooters. It requires “a tremendous amount of interoperability.” A new idea must be “backward compatible,” that is, consistent with other tools of, say, five years ago.

“How do we create, post facto, interoperability? By looking at all layers of the stack.” And just because you can move data from Point A to Point B does not mean they necessarily speak the same language. That calls for a fusion algorithm.

“A lot of artificial intelligence comes in,” Grayson said. “We are trying to streamline how humans interact with machines.” To succeed, Grayson said, “We don’t need a big bang, to produce in the thousands. We are happy to have thousands of little transitions for different users.”

“We want to let warfighters pose questions to a cottage industry, to commercial companies, who can throw an answer back over the fence,” he said. “We are trying to enable that ecosystem.” Grayson said DARPA would “love to work with SPIE to see how it applies across your research areas.” He has a doctorate in physics from the University of Rochester, where he specialized in quantum optics.