Air Force Chief Scientist Looking for Improvements in Automation
The coming years will bring many challenges to Air Force science and technology initiatives, with problems such as an uncertain budget and a dwindling talent pool posing a threat to innovation, a senior service leader said.
But as the Air Force goes toe to toe against new threats such as cyber-attacks and in contested environments, it will need intelligent, automated technologies that can reduce manual workloads and provide operators with trusted information, said Mica Endsley, the service’s chief scientist.
"The main benefits to automation are really in extending human capabilities," she said at a July 11 breakfast hosted by the Air Force Association. "It's not necessarily making decisions for people."
The threat environment will greatly change within the next 15 to 20 years, Endsley said. The Air Force has enjoyed air superiority throughout the past decades, but will be confronted with anti-access, area denial (A2/AD) situations that make it more difficult to carry out operations. Space also will likely be more competitive and congested, she added.
Cyberspace continues to be a significant source of potential vulnerabilities ranging from malicious insiders to attacks on the supply chain. Adversaries may also target command and control and intelligence, surveillance and reconnaissance capabilities in order to undermine the force’s ability to work as a team, she said.
Common to all of these areas is the need for improvements in automated technologies, Endsley said.
The 2013 Air Force Global Horizons report — which identifies the service’s technology gaps — stated that improvements in cybersecurity and connectivity present an opportunity to create remotely piloted aircraft (RPA) with a greater level of autonomy.
Endsley said human operators will continue to be heavily involved in the offensive operations of such aircraft. Right now, most RPAs are piloted by from the ground, she said. "Where we can start adding autonomy [is] for specific things, like navigation or communications. ... I think most of those are going to be incremental steps in various aspects."
In recent years, the U.S. military has greatly increased the amount of sensors that collect intelligence in the field, but there is still a gap between the massive amounts of data being taken in and the critical information to be extracted from that.
Autonomous systems could help analyze data, Endsley said. “It's transforming the data into useable information, and furthermore, making sure all of that data is integrated across the system."
Though various autonomous systems are already widely used by the military, a Defense Science Board report recently laid out areas where investment in autonomy could be useful, such as in anticipating failures and providing scenario assessments.
Part of the problem with current automated technology is that users don’t always understand the logic of a system or how it works, which sometimes leaves operators “out of the loop.”
"We need to make people be able to understand the assumptions and goals of the system, what it is the automation is projected to do, as well as what it's currently doing and how confident they should be in very specific situational and relevant data and outputs," Endsley said.
The service is also looking for autonomous systems with flexible modes, so that operators can use different levels of automation at different times.
Though Endsley spotlighted autonomy as a chief need for the Air Force, she laid out other gaps. The service wants improved precision navigation and timing systems to help airmen operate in A2/AD spaces, she said.
The Air Force plans to invest more funding into hypersonic and directed energy weapons. Tools to increase resilience during cyber-operations are also in demand. The service also wants to find better ways to support airmen with real-time physiological monitoring, she said.
Going into the future, the Air Force needs to look at how to better enable rapid innovation, prototyping and testing, Endsley said.
“I was down in Eglin [Air Force Base] a few weeks ago, and I saw where they're developing new munitions systems and new approaches,” she said. “They're taking it out on the range right there, testing them, getting rapid data, and coming back and being able to iterate their product and their ideas very rapidly. That's the kind of thing we need to be doing.”
Maintaining the technological edge could become more difficult as foreign nations increase their research-and-development funding and produce a higher number of graduates educated in the sciences.
The United States no longer commands as large a proportion of global R&D funding as it once did. It accounted for 42 percent of the money invested in new technologies in 2000, but is currently down to 29 percent, Endsley said. If China continues its current rate of growth, it will eclipse the United States by 2023.
Most U.S. research-and-development dollars will be spent outside the defense industry in areas such as pharmaceuticals, information technology and telecommunications, she added.
"We have to look very carefully and be very smart about our operations,” she said. “We think there is an opportunity to leverage the $1.4 trillion in global industry R&D. ... That's going to be essential to sustaining our current edge.”
The United States is producing fewer graduates with degrees in science, technology, engineering and mathematics (STEM), which may result in a shrinking pool of talent. This will lead to less development of revolutionary technologies, she said.
“With the increased number of STEM graduates abroad, we cannot necessarily assume that all the great science innovations are going to appear here. A lot of them are going to be appearing elsewhere," she said.