Until this week, U.S. Defense Department leaders had publicly described their technology race against China and Russia mostly as a bullet list of research priorities. Now a top research-and-engineering official has added detail about efforts to surmount key technical and physical challenges. At a Wednesday event put on by the National Defense Industry Association, Mary Miller, the assistant defense secretary for research and engineering, discussed directed-energy weapons, AI, quantum science, next-generation communications, and more.
It’s not just sharks with frickin’ lasers. Directed-energy weapons will be on jets, ships, tanks, and in space. All of the services, save the Coast Guard, are building direct energy prototypes and doing research in the area, said Miller, spending $661 million this fiscal year, and a planned total of $2.28 billion from 2019 to 2023.
“People will ask, why do you spend so much in directed energy?…We really need the services to understand and utilize” it, she said. The falling price of combined fiber lasers has opened the possibility of powerful and scalable lasers at cost, weight, and efficiency levels that would have been impossible a few decades ago. The department also sees near-term applications for high-powered microwave weapons, such Raytheon’s CHAMP electronics-frying drone.
But the department needs yet more, especially in terms of power, Miller said. “We are coming up on 100 kilowatts,” in capability she said. But “we have missions that require 300 kilowatt[s].” This may have been a reference to intercepting missiles from the air in boost phase. “I’ve even heard ‘2 megawatts’ out of the Navy,” she went on. (For reference, one megawatt will power nearly 1,000 homes.) “These are issues that are technical in nature, not in desire. We know what we want. We want more power.”
And yet in some cases, more power isn’t the key. “Research will also look at whether going to a one megawatt for a sea-based or ground-based laser makes sense,” she said. “There’s an aspect of physics called thermal blooming, which may make it such that more power just creates a plasma [electrified gas] in front of you and the power doesn’t get to the target.”
These are the sorts of issues the Department needs to resolve before committing to a direct energy program of record. “Prototyping and experimentation [will help the Department] figure all of these things out.”
Communications, Cyber, and Space
Uneven communication service has bedeviled the military for years. “We have small groups go out, become disconnected from the network, come back in, have to re-acquire the network and we have challenges with that,” Miller said. “It was one of those issues we were trying to work for our kids in Afghanistan. It’s not suitable where it stands today.”
On the future battlefield, every soldier, satellite, vehicle, and even nuclear weapon should be digitally linked, said Miller, echoing a goal put forward by top commanders. “We have this desire for joint, networked multi-domain battle,” she said; uniform and ubiquitous communications is a “critical enabler” of that.
“We are looking at how do we create a joint architecture so that we can all link the services underneath, so that, at the combatant command level, they have access and redundant access to comms at all layers,” she said. The goal should be communication that can’t be interrupted, because each party has many options or avenues to send the message or data. It’s an effort the military has tried before with limited success.
“If you need to send information and you have some nodes that go out of network? That’s okay. Much like a cell phone or an internet, you find another route. This is a very beautiful idea that the Army has tried for years. It’s a very difficult idea,” she said. A joint communications architecture, or standard, across service, platform, and domain, is a lot easier to implement if you are working in a more homogeneous data and network environment. That may explain part of the reason for the Pentagon’s insistence on a single provider for its massive Joint Enterprise Defense Infrastructure, or JEDI.
So the military is looking at how to create a joint architecture to “link the services underneath,” with an eye on commercial trends like the plummeting cost and size of low-earth-orbit satellites.
An additional concern is how to use old communication equipment in an environment defined by sophisticated jamming techniques. For years, the military and industry have been refining processes for taking radio functions normally embedded in hardware and recreating them in software. Such “software-defined radios” can move between frequences much more readily than traditional radios, hardening them against jamming. The trick is taking a sophisticated, cutting-edge technique and making it work in old equipment. “We are looking at how do you use software-defined radios to be backward compatible,” said Miller, saying that the experimentation is “in early days, but absolutely has to be done.”
On communications equipment, the military will spend $8.38 billion this year, and a planned total of $40.8 billion from 2019 to 2023.
On space, “we are looking at what new architectures to put in place,” meaning how to achieve the right mix of Defense Department and commercial satellites. To make it even harder to screw up communications between the soldiers, machines, the earth and the heavens, the Department is plugging a lot more money into light-based, or optical communications, whose line-of-sight beams are difficult to intercept.
The Department expects to spend $619 million on research and development for space modernization this year, and a total of $2.48 billion from 2019 to 2023.
One of the sleeper national security concerns of the next decade is the microelectronics market.
“The Chinese are being very forthcoming” about their ambitions to “dominate microelectronics globally by 2030” says Miller, highlighting U.S. estimates that China will spend $150 billion over the next 10 to 15 years. The Chinese government has forecast that China will own 70 percent of the global market for microelectronics by then. “This is something the United States Department of Defense is very concerned about,” said Miller. “We are significantly investing here ourselves to create a public-private partnership to bring microelectronics fabrication, research and design back to the U.S.” Research and design are the easy part. The fabrication is the big challenge, since U.S. industry outsourced that years ago to cheaper labor markets. It’s a trend the Defense Department is eager to reverse, at least somewhat. “We are leading the charge with industry who sees their future at risk.”
The Department plans to spend $42 million this year on microelectronics research, $2.2 billion between 2019 and 2023, according to a slide in Miller’s presentation.
Artificial intelligence and Autonomy
Much has been written about the department’s new plans to stand up a Joint Artificial intelligence Center. Miller cautioned against industry hype, saying that the most of what passes for “artificial intelligence” in marketing brochures today is actually just data science. The field is a long way from what the Defense Department really wants, expert systems that can help commanders and operators make much better decisions in real time. But the investment and the commitment is there.
She listed autonomy as a separate category with its own set of challenges rooted in the interface between humans and smarter machines. Getting humans to trust their interactions with highly autonomous systems is a new and rapidly advancing field. In private and commercial settings, there’s plenty of low-stakes experimentation. The Defense Department use case is rather different.
The department has a sense of what it wants from artificial intelligence in terms of refining logistics and expediting decision making, with autonomy, the policies that will guide their use remain in flux. “We will have worked out the policy… by the time you get it,” she promised on autonomy.
The department will spend $1.9 billion on autonomy research this year, and some $10.3 billion between 2019 and 2023.
Similarly, Miller noted that quantum science research is still in its infancy and “certainly overhyped,” especially in terms of scalable quantum computation — that is, making a computer that manipulates qubits rather than binary bits. But the Defense Department has already found some near-term uses for the science, in precision navigation and timing, for which it currently relies mostly on the Global Positioning System.
In the last few years, accelerometers, atomic clocks and gyros have improved a thousand-fold, Miller said. The Department, she said, is also exploring “crypto-modernization to protect against the threats of any future quantum computer,” addressing a common, if perhaps misplaced, concern among information security geeks.
This year, the department will spend $96 million on the field, with a planned $565 million between 2019 and 2023.