Location, Location: One of IoT’s Greatest Obstacles

As the Internet of Things (IoT) develops and grows more connected, there will be an explosion of data points to track, analyze, and record. Naturally, this will bring issues in lack of hardware and software to adequately respond in real-time to situations.

Location is the biggest issue for IoT right now. By itself, location data is incredibly multi-dimensional and multi-faceted. Information about time, nearby devices, current device, and more surround this type of data, and it’s generated every single day.

A New Data Generator

In the past, the solution that IoT professionals implemented involves having devices report to a hub in a centralized network. But the problem presented itself almost immediately with this approach: this solution isn’t suitable for mobile devices.

As 5G continues gaining traction and support, we expect to see it across all smart cities, incorporated into roadways, cars, buildings, devices, and much more. This super-fast network connectivity affords us real-time updates, but it also generates a huge amount of data.

“We used to think about IoT devices as static,” Usman Khan, who is a professor of electrical and computer engineering at Tufts University. “But we’re moving to this Internet of Moving Things.” Khan is absolutely correct with his insight; we need to incorporate all mobile and static devices into one network that can be reliably searched, audited, and analyzed in one go.

GPS’s Limitations

Khan is working with other researchers at Carnegie Mellon University and Tufts, based outside of Boston’s city limits. When tackling this problem, the group realized that “we cannot possibly implement all of these with GPS. It’s too costly. Even if it’s only 10 cents each, it’s expensive,” says Khan.

Historically, GPS also has issues working inside without a Wi-Fi connection. Connecting entire warehouses to Wi-Fi, even that unreachable spot in the dead center of the floor, is highly unrealistic.

This type of non-precision is utterly useless when it’s necessary to track the real-time location of multiple devices. Forget about figuring out object relationships to each other; without accurate location data, almost all location data generated is unusable.

A Better Proposal

The group realized that not every device needs GPS; a few devices could have GPS as anchors, like devices that remain static. “The idea is, none of the devices need to talk to the anchor directly,” says Khan.

In fact, only a few devices would really need to communicate with the anchor. Using that data, the closest devices could self-locate themselves, allowing others nearby to locate themselves also.

This is theoretically a self-sustaining model for device localization that doesn’t come without some issues, though. Since the anchoring device model uses linear equations to help devices figure out where they are, the limitation comes from the linear equations themselves. These equations are fast but not as accurate as non-linear equations, which cell phones use to triangulate position between three cell towers.

Bluetooth to the Rescue?

Kathleen Philips is a program director at Imec, a Belgian electronics tech company that recently unveiled a Bluetooth-based solution. Early results were very positive: devices within 30 centimeters can be discovered.

“It’s not just about localization, it’s about spacing,” Philips says. Right now, the Bluetooth implementation works for devices that are moving below the speed of a bike. And it can easily juggle several hundred devices before it involuntarily drops some from its network.

Although this solution isn’t perfect yet either, it’s great to see that researchers all around the world are working hard to tackle this IoT problem. Perhaps the best solution will be a mixture of the Bluetooth localization and GPS implementation. What do you think could help IoT track devices better and more accurately?