Lt. Col. Nathan Wiedenman, AVM's deputy program manager, said one huge reason programs tend to exceed their budgets is that they take so long to complete. And the reason they take too long is they have to be close to fully-assembled before designers can see whether they work properly.
"So how do we currently design these big, complex systems? We do it the same way we've been doing it for over 50 years," he said. "We break down the systems we need, typically along engineering disciplinary lines. This is a power system; this is a thermal system; this is a drive system, for example. And we make sure that all the parts are the best possible for their individual tasks. And then we put it all together, and we build it. After we build it, we test it to see if it works the way we expected. And of course, invariably, it doesn't, so we have to go back and redesign, rebuild, retest and so on. That takes a lot of time and a lot of money to iterate like this."
That process has been what it's been, Wiedenman said, by necessity. Up until now, there hasn't been a way to know how the millions of components of a high-tech weapons system would work together without actually building one.
New concept to develop products
That's the challenge DARPA is trying to solve. They're working to build a set of software tools that would let the Pentagon and its contractors develop products with a concept called "correct-by-construction." Instead of taking up the resources of a factory floor to build a prototype and successive test models, a product works the way it's supposed to the first time it rolls off the assemble line.