When GPS is silent, the military relies on this system to take its place
For most of us, GPS has made getting lost feel like a thing of the past. True, a mobile phone may occasionally lose its signal — in dense cities, for example, or tunnels, or steep, rocky terrain. Yet as long as the phone has some juice, chances are good it can pin down its owner’s location to ten or so meters while pointing the way to a destination. And even when it can’t, close is often good enough.
But for the military services, the stakes are much higher. Precisely locating an aircraft, naval vessel, ground vehicle, guided projectile, or a squad member on foot from minute to minute, or even second to second, can be essential to completing a critical mission safely.
The mere risk of losing GPS in urban areas or rocky canyons is reason enough for concern. But now those worries are compounded by the threat of a GPS signal being intentionally removed or distorted by adversaries as part of a hostile action, often referred to as jamming or spoofing. “The military has become reliant on knowing exactly where its millions of personnel and vehicles are at all times all over the globe,” says Kevin Betts, the Position, Navigation and Timing (PNT) Director at Leidos. “Suddenly, having to operate without that capability would significantly impede its movement and strike capabilities.”
Location data without GPS
Is there a way for the military to maintain positioning and navigation awareness inside the bubble of a so-called “GPS-denied environment”? The answer is now yes--thanks to a suite of technologies developed by Leidos that can instantly step in when GPS steps out. This is part of an overall solution the DoD calls Assured PNT.
Leidos’ solution typically relies on two basic types of sensors to generate a reliable position fix. One type is any of a range of external sensors that can measure what’s around users or vehicles, such as visual information, the earth’s magnetic field, or signals from radio towers. The other type consists of inertial sensors, which can sense relative motion. Any time the first type of sensor can provide enough information to offer even a temporary fix, such as by recognizing surrounding scenery, an inertial sensor can then track how the position is changing until a new fix is available.
That sounds simple — but it’s not. Extracting reliable position fixes from an unpredictable, shifting combination of sensors with signals that may come and go, and that may range significantly in accuracy, is a massively complex task.
Developing ADEPT
Leidos needed to develop a software platform that could wring a fix from whatever sensors happened to be available from moment to moment on a mission, whether on the ground, in the air, or in or under the water. And the software had to run on whatever type of computer processor is on hand, whether a powerful shipboard server or a warrior’s handheld radio.
The result is Assured Data Engine for Positioning and Timing (ADEPT). Originally developed as part of an ongoing DARPA program called All Source Positioning and Navigation (ASPN), ADEPT is already proving in tests that it can meet the formidable challenge of navigation in GPS-denied environments during almost any type of mission. “ADEPT does whatever it takes to extract positioning information from any sensor data that is available,” says Leidos PNT Program Manager Michael Turbe, who helped develop the system.
Visual data is often the key to getting a position fix when GPS is off the table. For one thing, it’s widely available — affordable, compact, high-resolution cameras are now routinely mounted on everything from aircraft wings to soldier helmets. These cameras can capture nearby terrain features or manmade structures that can then be matched to digital maps and satellite imagery.
Of course, visual clues to a location are hard to come by for warriors in dense woods, vessels far out at sea, submarines, or aircraft in or above clouds. That’s where other sensors come into play. Odometers and compasses, signals from cell phone towers and non-GPS satellites, barometer readings that indicate changing terrain height, even minute variations in the earth’s gravitational pull, all those data sources, and many more can play a role. And whether it’s a submarine a hundred meters down, a guided missile streaking over mountains, or a Humvee moving along a canyon floor, there are usually at least a few sensors at hand to pick up non-visual clues to terrain and location.
ADEPT’s first job is to grab the different streams of available sensor data and translate each one into a form that the system can make use of, drawing on a large, constantly updated library of sensor software drivers developed by the ADEPT team. Then the data is sent through an array of sophisticated electronic filters, in some cases backed by artificial intelligence, to derive useful location clues. For example, the view from a camera mounted on a ground vehicle, such as a tall building on one side of a road or a small hill on the other, doesn’t look anything like the overhead view of the same terrain on a satellite map. But machine learning software based on robot-vision research can figure out what those features would look like on a satellite map, in order to match the images up.
That’s the relatively easy part for ADEPT. What’s harder is to take those different sensor-derived clues, determine which are the most informative and reliable, and then blend them together to pull out the most precise, accurate fix possible, all in a fraction of a second. “There’s a lot of cutting-edge mathematics that goes on inside to make that happen,” notes Turbe. The result is what the team calls a “sensor fusion engine,” and it’s the very heart of ADEPT.
Even better, it means that the sensors carried by a single warrior, aircraft, or vessel don’t have to kick in all the data needed for a fix. ADEPT can pull the required information together in bits and pieces from sensors spread out over two or more people and vehicles operating as a group, by taking into account the distances and angles between the various sensors. When enough of this scattered data is woven together to provide a trusted location for even just one person or vehicle in a group, everyone else’s fix can be instantly derived from it, forming what’s known as “collaborative navigation.”
GPS remains the ultimate in precise navigation, says Betts. But he notes that ADEPT is proving itself a solid substitute, one that could become critical in a wide range of situations. “There’s usually nothing more useful on a mission than knowing where you are,” he says.
