The test was a smashing success — but not at all a surprise.
Off the coast of Hawaii, a Standard Missile-3 Block IIA interceptor launched from the guided missile destroyer USS John Finn, rocketed outside Earth’s atmosphere and destroyed an intercontinental ballistic missile target.
It was the first test of its kind in the physical world. But in the digital world, it had already happened – many, many times.
Using artificial intelligence, machine learning, modeling and simulation, engineers at Raytheon Missiles & Defense digitally put the interceptor through its paces time and again in the months leading to the November 2020 test.
“We knew we could do it long before we proved it,” said Wes Kremer, president of Raytheon Missiles & Defense, a Raytheon Technologies business. “That is the kind of powerful insight you gain from modeling and simulation.”
The simulations leading to the real-world test were just one way Raytheon Missiles & Defense is using artificial intelligence and machine learning to assess data from its own products, predict their performance in real-world scenarios and help customers make better purchasing decisions.
In the SM-3 test, for example, the team gathered data from previous tests and fed them into a model and simulation.
“We do the high-fidelity predictions that tell us how the missile’s going to perform throughout flight, all the way down to the exact point on the target the interceptor will hit — and we were right,” said Doug Fiehler, part of the Raytheon Missiles & Defense flight test team at the Missile Defense Integration and Operations Center at Schriever Air Force Base in Colorado Springs, Colorado.
A fact-finding mission
The process, as it often does, starts with listening; a team from Raytheon Missiles & Defense meets with a customer to hear what they want.
“What we’re really after is having that discussion with the customers to get them to open up and say what they really need, what keeps them awake at night, and what we can do at Raytheon to help solve some of their problems,” said Bob Fitzpatrick, vice president of Requirements and Capabilities at Raytheon Missiles & Defense.
Modeling involves examining those needs with a team of experts, what they call “the innovators,” to determine physics and design parameters that are then put into action with simulations.
Try before buying
Modeling and simulation have two enormous advantages over building and assessing actual hardware: It’s much cheaper, and it’s far faster, allowing the business to respond rapidly to any changes the customer requests.
“AI and machine learning are critical to the company’s modeling and simulation framework,” Fitzpatrick said. “Every new flight test and every new experiment looks a little different than the last one and that means there’s a lot of rapid learning happening for the company and its customers.”
For example, the hypersonic threat is evolving and it’s not entirely clear which weapons adversary nations have – or what they can do. Modeling and simulation allows engineers to run thousands of variations to see what works best.
“We visually take them step by step through a conflict and show them what it would mean to buy or upgrade a system,” Fitzpatrick said. “We offer them options they may not realize they have.”
The company likens this approach to a football coach studying old game tapes to see how different players with different capabilities impact a game for better or worse and how they can adjust their response to either exploit a team’s weakness or counter their strengths. Likewise, the company runs multiple plays and manipulates various systems to determine the best way to respond.
“No one can afford a perfect missile, no one can afford a perfect sensor, so you really must get them playing together in a synthetic battlefield and see how they tie together in order to maximize a customer’s investment to ensure they’re not overspending for perfection,” Fitzpatrick said.
A system of systems
This new generation of software is helping ensure troops are ready — equipped with modern systems that are connected and can conduct a full range of operations.
“We cover the entire kill chain,” Fitzpatrick said, using a military term for all the systems in an operation such as an intercept or counterattack. “From sensing, communications, command and control, and launching a weapon of our own, we digitally simulate all of these interactions so we can quantify the effectiveness of our system of systems.”
The company uses modeling and simulation to show how new systems can integrate with existing systems – foreign or domestic.
“It’s not just about getting every piece of the puzzle perfect,” Kremer said. “It’s about the whole puzzle and helping the warfighter understand how they can better use current and future systems and capabilities to gain the advantage on the battlefield.”