The Defense Department over the last decade has built up an inventory of billions of dollars worth of spy aircraft and battlefield sensors. Those systems create avalanches of data that clog military information networks and overwhelm analysts.
Intelligence experts say the military is drowning in data but not able to convert that information into intelligible reports that break it down and analyze it.
“The challenge for users of intelligence is that all the different types of information come in a stove-piped manner,” says Michael W. Isherwood, a defense analyst and former Air Force fighter pilot.
Intelligence feeds include electronic signals, satellite imagery, moving-target data and full-motion video. “How do you integrate this into a clear picture?” Isherwood asks. “That is one of the enduring challenges in the ISR [intelligence, surveillance and reconnaissance] arena for all the services.”
Isherwood, the author of a Mitchell Institute white paper, titled, “Layering ISR Forces,” cautions that success in future operations hinges on “timely, astute combinations of ISR resources.”
The Pentagon would be wise to shift its future investments from sensors to data-analysis tools, he says.
“The awareness gained from integrated, multi-source intelligence data is of supreme value,” says Isherwood.
In actual combat, a coherent picture of the battlefield is not a “routine event,” he says. “Coalition forces in Afghanistan have suffered losses when they were surprised by a much larger insurgent force not detected in time by ISR assets.”
Military drone operators amass untold amounts of data that never is fully analyzed because it is simply too much, Isherwood says.
In the Air Force alone, the buildup of data collectors has been dramatic. While its inventory of fighter, bomber, tanker and transport aircraft shrank by 11 percent over the past decade, ISR platforms — primarily unmanned air vehicles — increased by nearly 300 percent, says Isherwood.
Air Force leaders have recognized this problem and recently decided to cut its future purchases of Reaper drones in half — from 48 to 24 — because there is not enough manpower to operate and process the data from more aircraft. “It didn’t make sense to have the production out that far ahead of our ability to actually do the processing and exploitation and dissemination function,” Deputy Assistant Secretary of the Air Force for Budget Marilyn Thomas says at a February news conference.
The military services have funded programs to develop software algorithms to automate data analysis, but no silver bullet has emerged.
“Industry is working on tools so you can pull a Google Earth image and incorporate the SIGINT [signals intelligence], the MTI [moving target indicator], visual imagery, full-motion video,” Isherwood says.
What the military needs is a “decathlete analyst” that can process multiple feeds, versus an operator for each type of data, he says. Defense Department leaders understand the problem, but the “acquisition community now needs to take that and translate it into systems” that tackle this challenge.
The Air Force is “really good at building an airplane,” Isherwood says. But he has yet to see a comparable requirements document or request for technology that meshes all the sensors, he adds. “They go after it piecemeal.”
The information deluge problem also is exacerbated by the military’s organizational silos that zealously protect their data.
“It’s hard to get the community to plug their sensors in,” says Gregory G. Wenzel, vice president of advanced enterprise integration at Booz Allen Hamilton.
The so-called “PED” process — processing, exploitation and dissemination — has been a long-standing challenge, he says. “It’s a really hard problem.”
Automated analysis tools for video feeds are gradually entering the market, Wenzel says. The National Football League has developed software to search video archives that some defense contractors are using as a model.
One of the more promising systems that could help military ISR operators manage data more efficiently is the DI2E, or defense intelligence information enterprise, says Wenzel. The entire Defense Department and intelligence community will be able to share information, he says. The DI2E is a cloud-based system that draws data from many sensors and databases.
Technologies such as DI2E are part of a larger trend toward networking sources of information, says Richard Sterk, senior aerospace and defense analyst at Forecast International. “There’s still too many stand alone legacy systems.”
Regardless of advances in technology, he says, a larger conundrum for the military is figuring out how to manage information so commanders and troops in the field don’t become overwhelmed. “They have to sort out how much information is enough,” says Sterk.
The Office of Naval Research and the Marine Corps have been experimenting with another approach to analyzing data known as “semantic wiki.”
It solves the “intelligence fusion” problem, says George Eanes, vice president of Modus Operandi, a small firm that developed the wiki tool.
It’s a rather simple approach. “If I’m looking for something of interest, like a white van, I can search across all the data stores that I have access to,” Eanes says. “It presents it in a wiki format. … It’s a really good tool for pulling the data in from multiple sources and present it in one convenient application.”
Semantic wiki can search video, human intelligence reports and satellite imagery. Streaming video could be added in the future, he says. The company has spent the past three to four years working on this technology under several small business innovation research contracts worth about $5 million, says Eanes.
“There has been a cultural shift within the Defense Department toward more desire to share information,” Eanes says. “First they thought the solution was to bring everything into a single database. But that proved impractical. There is too much data,” he says. “Now they’re looking at other solutions. You keep the data where it resides. You access only the data you need.”
Former Marine Corps intelligence analyst Tony Barrett, who is now at Modus Operandi, says that during his time on active duty, his team was overwhelmed by data. He would have liked to have had software to scan unstructured data and provide relevant information, based on queries the analyst sets up, he says. “That frees up the analyst to do due diligence rather than extended periods of research,” he adds. “In Iraq, I had individual analysts that all they did was scan reports and find which ones were relevant. … Research is extremely frustrating. I would rather my guys spent more time thinking.”
Because of the data overload, “What you end up doing is taking your smartest Marines who would be your biggest help in problem solving to work on your system’s problems,” Barrett says. “I had my smartest guys always be the principal researchers because I was more confident they would be able to discover more data than less talented analysts.”
ISR experts also worry that the military has become addicted to full-motion video, at the expense of other intelligence disciplines that might gradually disappear as the number of skilled operators declines. Video imagery is the most “readily understood” intelligence, says Isherwood.
For the Iraq and Afghanistan wars, full-motion video provided by aerial sensors was the preferred form of surveillance. But for other combat scenarios in the future, Isherwood says, the military might need to rely on other types of data such as signals intelligence (collection of electronic intercepts or emissions), moving target indicator data (Doppler shifts of moving objects to detect and track targets), radar imagery; and measurement and signals intelligence (combines radar, laser, optical, infrared, acoustic, electromagnetic and atmospheric mediums to identify objects). There is also “cyber-intelligence,” a new discipline that is based on electronic-warfare techniques, says Isherwood.
“Full motion video is what everybody wants,” says Chief of Naval Operations Adm. Jonathan Greenert. “A still picture is good but you still have to send it back, develop it quickly,” he says.
Access to full-motion video, however, might not be feasible in every conflict. “Not all fights will be in the desert,” says Mel French, vice president of development at Telephonics, a supplier of military sensors and electronic warfare systems.
The unmanned aircraft-mounted sensors that are favored today might not work in other environments. “The second you introduce rain to any of those systems, the range goes down, it limits utility,” says French.
“We need to think of where else we are going to go,” he says. “Possibly places where we need foliage penetration. That’s a hard problem to solve.”
The full-motion video soda straw view works when the area is not being defended by adversaries who can shoot down surveillance aircraft, he says.
In instances when ISR assets might be in danger and rather kept at standoff ranges, the military will need analysts who can discern other forms of data such as synthetic aperture radar images, French says. Some field commanders might complain that they “don’t understand the [SAR] shadows,” he says. They might not realize that video camera pictures can’t be obtained from 200 miles away. Images such as SAR require a trained eye.
As to whether there will be a time when analysts will be able to produce “actionable” intelligence, French says there are no easy answers.
“It’s one of those problems that will require years of investments and focus,” he says. “We fielded a lot of Band-Aids. Now it’s getting back to rationalizing what we fielded.”