It’s hard to find things sometimes. Ask anyone who’s lost their keys.
Now imagine trying to locate thousands of sets of keys that are wildly roaming around. This dada-esque scenario is precisely what Internet of Things (IoT) networks will soon have to deal with—keeping track of dozens, hundreds, or thousands of devices as they move about.
Typically, the solution has been to have devices report back to some sort of hub in a centralized network. And that works well if the devices aren’t particularly mobile—stationary sensors tracking weather changes, for example—or if there aren’t very many of them. But the advent of 5G means there may soon be too many devices for a centralized hub to reliably track.
That’s why researchers are looking for new ways to keep track of them all. Some are focusing on the hardware side of the problem, others are looking at the software side, but the goal is the same: Know where every device in the system is at all times.
“We used to think about IoT devices as static,” says Usman Khan, a professor of electrical and computer engineering at Tufts University. “But we’re moving to this Internet of Moving Things.”
That’s why Khan, along with several other researchers at Tufts University and Carnegie Mellon University, has been thinking about new ways for devices within IoT networks to communicate their locations.
Imagine a fully automated factory floor. Perhaps there are a thousand robotic arms and conveyor belts and forklifts and gadgets and gizmos working to assemble, for example, an autonomous car. Some of those gadgets will be stationary, but plenty of others will be mobile, and it would certainly be ideal if they avoided crashing into one another, or their stationary counterparts.
Or, imagine a hospital with all sorts of smart sensors attached to gurneys, being wheeled around the various floors and wards, or luggage moving through an airport. “We cannot possibly implement all of these with GPS,” says Khan. “It’s too costly. Even if it’s only 10 cents each, it’s expensive.”
Besides, GPS—a common localization solution—doesn’t work indoors, since you need a clear line of sight to satellites in orbit in order for it to work. And GPS doesn’t offer the precision needed for devices moving about on a factory floor or in a hospital.
That’s what sparked a realization for the group of researchers, who published a paper on the subject in May: Not every device needs GPS. Instead, you only need a handful of devices to function as GPS anchors. These anchors, as the name implies, remain fixed.
“The idea is, none of the devices need to talk to the anchor directly,” says Khan. Instead, only a handful of devices—the closest ones—communicate with the anchor, triangulating their position by measuring their distance from the anchor as well as from other nearby devices.
From there, the next closest devices can locate themselves based on where the devices closest to the anchors know they are, and so on. Eventually, Khan explains, you’ll find devices that are nowhere near the anchor, but still know exactly where they are, based on location data from their neighbors.
There are some limitations to this strategy, however. It relies on using linear equations to locate devices, and can only work in a pre-defined space. The strategy couldn’t replace, for example, how a smartphone determines its location by using three cell towers. One tower is enough for the phone to know it’s somewhere on a circle with a radius determined by how long it took a signal to reach the tower. With two towers, the phone knows it’s at one of two locations determined by the intersections of the circles around each tower. But it takes a third tower to provide the last bit of information needed to determine in which of those two spots the phone is located.
That solution uses nonlinear equations, which are slower to solve, but work regardless of whether the phone is somewhere between the three cell towers, or outside of that area. The solution Khan and his co-researchers developed uses linear equations, but if a device wanders beyond the boundaries established by the anchors, it would not be able to figure out where it was.
Imec recently announced a Bluetooth-based localization solution that can pinpoint a device within 30 centimeters. Imec’s work is unrelated to the work of Khan and his fellow researchers, but shows that many minds are working to find better localization solutions.
As Kathleen Philips, a program director at imec, explains, not only is the need greater for locating more devices—we also need to locate them more precisely than ever.
“It’s not just about localization, it’s about spacing,” says Philips. For example, a smart lock should only unlock if the homeowner’s smartphone is right on the doorstep, not five meters away, still walking up to the door. In hospitals, doctors should only be able to pull up confidential patient information on a tablet if they’re right next to the patient’s bed, to prevent just anyone from accessing it.
Philips says imec’s Bluetooth solution works well for anything traveling at about the speed of a bike, or slower. It can also handle a few hundred devices communicating before individual devices would begin to fail. There are still some details to iron out—for example, making sure that a smart lock can tell the difference between when the smartphone is indoors versus outdoors, so that the front door doesn’t unlock each time you walk past it while already inside.
With 5G, more devices than ever will be deployed to work with little to no human supervision. The better they can talk to one another, and know where all of their fellow devices are, the easier it will be for them to carry out their work.
Michael Koziol is an associate editor at IEEE Spectrum where he covers everything telecommunications. He graduated from Seattle University with bachelor's degrees in English and physics, and earned his master's degree in science journalism from New York University.