JUST IN: Navy shares details on new strategy for “intelligent autonomous systems”
MQ-4C Triton
Photo by Northrop Grumman
Following the publication of the Department of the Navy’s Unmanned Campaign Framework earlier this year, the Marine Service recently unveiled a new complementary strategy that focuses on intelligent autonomous systems.
The March Framework – which has an overarching vision of making unmanned systems a trustworthy and sustainable part of the naval forces – emphasized platform capabilities and identified a variety of technologies required for the rapid development and deployment of new systems.
Jason Stack, technical director of the Office of Naval Research, said the Navy’s new complementary science and technology strategy, Intelligent Autonomous Systems, will be key to unlocking the potential of unmanned platforms.
“IAS will be a cornerstone of future maritime power,” he said on July 27th during his speech at the Association for Unmanned Vehicle Systems International’s Unmanned Systems Defense Conference.
IAS is the confluence of autonomy with unmanned systems and artificial intelligence, according to the strategy published in July.
“To prevail in the IAS competition, we have to act quickly and concentrate on the most effective results,” says the strategy. Goals include establishing a smart autonomous vision for the Navy, providing an investment management framework for funding, manpower and infrastructure; and accelerating the development, operationalization and introduction of IAS systems.
The focus areas include skills, people and processes, and partnerships, the document says.
At the most basic level, adding autonomy to an unmanned system makes it adaptable, Stack said.
“When we make machines smarter, they ultimately get the ability to adapt,” he said. “When we think about what research to do, what to experiment with [concepts of operations], how [we] As a developer community needs to come together and provide more options and alternatives to DoD, what matters is understanding how to make our unmanned systems more adaptable. ”
Systems must be able to adapt to unstructured and unpredictable scenarios at machine speed while feeding in overwhelming amounts of data, he said.
“When you put these three things together and you have machines that can work effectively with people, where the machines can adapt in that way, [then] This is a truly revolutionary issue, a truly revolutionary issue in relation to something the world has never seen before, ”he said. “Precisely because of this … we claim that IAS and these kinds of skills combined with our people will be the cornerstone of future maritime power.”
Stack pointed out that the Navy must “innovate disruptively” on its way forward.
“We need to rediscover our ability to innovate in a highly disruptive way,” he said. “We have gone through history many times where we absolutely excelled at disruptive innovation and it is up to us to rediscover our ability to do so.”
The three attributes of disruptive innovation include early and frequent experimentation, encouraging bottom-up opportunities and technological discovery, and prioritizing action, learning and feedback in processes, he said.
“This is the mantra of the entire Silicon Valley and small and lean startups, so to speak,” he said. “You prioritize measures, learning and feedback beyond planning in order to minimize risks. … One of the main characteristics of disruptive innovation is undressing, learning while you walk. ”
The Pentagon often struggles to strike the right balance when pursuing new systems, Stack noted. “When you think of the old adage of measuring twice and cutting once … in DoD we often measure 10 or 15 times before cutting,” he said.
When building a nuclear reactor for a submarine, this is probably the right approach, he said. On the other hand, a marine or sailor testing a small quadcopter enables a different level of risk management.
“It is up to us to come back and rediscover our ability to make up for this,” he said.
Subjects: Robotics and Autonomous Systems, Navy News
Comments are closed.