In the night sky somewhere over Syria, a MQ-9 Reaper orbits in near-silence. All but invisible even in daylight, the 36-foot drone has absolutely no natural enemies after dark.
Seven thousand miles away, midday, at an undisclosed location somewhere in Nevada, the drone operations team actively tracks a terrorist cell.
As the pilot holds a steady orbit, the information streams in. A single sensor suite floods the satellite downlink. During the drone’s 14-hour loiter time, the ARGUS-IS imaging system can stream up to a million terabytes of data and record 5,000 hours of HD (High Definition) footage using 368 Focal Plane Array imagers.
- Syracuse University - M.S. in Applied Data Science: GRE Waivers available
- SMU - Master of Science in Data Science - Bachelor's Degree Required.
- UC Berkeley - Master of Information and Data Science Online - Bachelor's Degree Required.
- Syracuse University - Master of Information Management Online
- Maryville University - Master of Science in Business Data Analytics
- Villanova Business - Master's in Analytics and Study Data Mining, Predictive Analytics Online
The system can independently:
- Create up to 100 separate video windows
- Detect and track moving vehicles and foot soldiers
- Reference a forensic archive of past mission data
- Generate 3-D models of contacts
Sun Tzu said that the key to striking and conquering is foreknowledge. Today, the key to foreknowledge for military and counter-terrorist forces are master’s-educated data scientists.
Data is Critical to Viewing the Modern Battlefield
The capabilities of the ARGUS-IS and other modern sensor suites is formidable to the point of black magic. But, according to defense industry publication Defensetech, the military is having trouble finding people to mine all the data produced by those platforms. The Department of Defense has reached out to such unlikely companies as National Geographic and ESPN to understand better ways to manage the volume of information coming in.
Automation is one solution to dealing with the influx. As with the ARGUS-IS, data scientists are using machine intelligence solutions to comb the data streams and identify possible targets to be reviewed by a human analyst. In 2003, fledgling algorithms could take as long as 60 seconds to identify a vehicular target. By 2016, acquisition happens in real-time.
These algorithms are not only used to acquire targets on the conventional battlefield. On the human battleground of counter-terrorism, data scientists are exploring modeling techniques to understand:
- Patterns behind suicide bombings and IED (Improvised Explosive Device) placement
- Explosion simulations for various types of explosives
- Logistical challenges for terrorist forces with the goal of cutting supplies and access to explosives experts
- Blast signature analysis
Military data scientists have to deal with a challenge not common to their civilian counterparts, which is dealing with intentional obfuscation and interference of data collection. Spoofing, jamming, and deception are routine parts of military and terrorist encounters. Highly-trained data scientists are continually working to develop algorithms to see through such deception.
Conventional forces are embracing the Internet of Things, which allows previously “dumb” devices like tanks and rifles to have embedded logic and networking features, further increasing the flood of available data coming in. Counter-terrorist forces are also tapping into civilian sources of data including security cameras, cellular networks, and public records databases to improve access to information and more effectively track adversaries.
For both military and counter-terrorist forces, the challenge for data science is to blend automated detection with artful interpretation by professional analysts to produce actionable intelligence.
Detection and Interpretation are Worthless Without Action
Getting the data from drones and other sensor platforms and analyzing it is only half the battle, though. Transferring actionable intelligence to soldiers and agents in the field fast enough for it to be useful is the other half.
The networking challenges of establishing connectivity with field operatives and troops is a thorny problem in and of itself, but it pales in comparison to the difficulty of presenting data quickly and understandably in high-stress situations.
Big Data is a double-edged sword for troops in the field. The insights that can be gleaned from the information can give them a critical advantage over their opponents, but the fire hose of data can obscure vital information beneath a stream of less critical updates.
Data scientists face a significant challenge in designing systems that can turn a data stream into contextual information that troops can use to enhance situational awareness. Many officers are frustrated that sensor and intelligence gathering platforms have significantly expanded capabilities in recent years, but that interface and reporting design hasn’t kept up.
A 2012 Navy report titled “Lightening the Information Load” quoted a Marine officer in Afghanistan as saying, “…90% of the information requested was already in the system, but the person asking didn’t know where to look or how to extract it.”
Under extreme stress, as is common on battlefields, humans become easily task-saturated and have difficulty focusing on complex interfaces. The Navy report calls for data scientists to take advantage of machine intelligence to distill incoming data to highly portable and understandable “app” solutions so as to deliver short bursts of unambiguous and actionable intelligence to end-users in the field.
Moving Pieces On the Board: Tracking Forces with the Internet of Things
On the battlefield, information about the enemy is only slightly more important than information about one’s own forces. A spate of friendly-fire, or so-called “blue-on-blue” incidents during the Persian Gulf and Iraq Wars demonstrated the danger of not keeping attack forces fully informed about the location and capabilities of one another.
The response to these problems in the U.S. military was a program called “Blue Force Tracker.” Again leaning on the Internet of Things paradigm, Blue Force Tracker (BFT) embeds a GPS (Global Positioning System) receiver with a satellite transceiver and software in tanks and other military vehicles. The vehicle’s position and other status information are uplinked to military communications satellites and then linked in to data from other vehicles and systems to provide a comprehensive real-time picture of all assets in the area. A glance at a screen can inform commanders or other vehicle crews of the position of other BFT-enabled units, even if communication has not been established.
In addition to reducing incidents of unintentional fratricide, BFT allows commanders to view force deployments and optimize routes based on terrain and tactical plans.
Feeding the Machine: Military Logistics Take Advantage of Big Data
Civilian logistics operations are a big deal to global commerce, but military logistics are life and death to the troops. What is now commonly known as logistics, first originated as a science that dealt with military supplies and supply lines. And with modern battles fought, more often than not, on the other side of the world, getting Reaper parts and fuel to forward airstrips is more important than ever.
Today, data scientists are working to optimize military supply chain management just like their civilian counterparts at FedEx.
The Department of Defense is even taking a page from FedEx and Wal-Mart, instituting such practices as chipping assault rifles with RFID (Radio Frequency Identification) devices to assist with inventory management.
The military is also using more prosaic applications of Big Data to deal with budget limitations. The DoD absorbs almost one-third of the national budget each year. It manages a massive physical presence scattered around the world and employs the largest workforce in the world, 3.2 million strong. Programs like the Automated Energy Audit, which combs through building environmental control data to search for up to 2,000 different optimizations with the goal of reducing energy utilization, can have a major impact.