As an electrical engineer and professor of aerospace engineering, I have intimate knowledge of the core technologies that go into GPS: signaling, RF propagation, and spread-spectrum coding among them. Yet I still see GPS as a miracle — that something as complex as it is, with so many interacting parts, works as reliably and precisely as it does. It enables us to navigate nearly all over the earth, at high update rates, with ten-meter precision, thanks to the enormous sponsorship of the US Department of Defense.
When the GPS constellation was first launched, its promoters within the government could not even foresee the need for civilian receivers. Now we have GPS chips in every cellphone, innumerable apps that enhance our daily lives, and myriad unseen uses that enhance the world of commerce.
24 GPS satellite constellation and visible satellites from Golden, CO (via Wikimedia)
But GPS has crucial limitations: it doesn’t work well in cities, it doesn’t work indoors, and it doesn’t work underground. Yet this is where most people, and most robots, live and work these days. These realms, the human-made world, are usefully referred to as “the built environment.” Have you ever emerged from the subway in New York City and tried to use GPS to figure out which side of the street you’re on? Have you ever tried to connect to an Uber driver, convinced you were both in the right spot, when in reality you were on opposite sides of the block?
GPS doesn’t work in the built environment because it depends on multiple lines of sight to satellites, which is lines of sight to the sky. As you read this, look up and see how much sky you can see — whatever proportion you can see is how well GPS will work, and the answer is likely zero. Software tricks (like getting the coordinates off a nearby wifi router) can mitigate some of these problems, but not with the kind of precision we expect from our location apps. In cities, multipath reflections of GPS signals can degrade the accuracy to hundreds of meters. The GPS display in your car often masks this effect because it “snaps” you to the nearest road….hopefully the one you’re driving on. Various other clever software fixes are helping to improve the situation, but they always start with the same basic signal limitations.
How will we navigate in the built environment? The answer is not going to be a single system like GPS that covers the entire earth — the nature of the built environment is that it is fragmented and local, broken into innumerable small geometries. Subways run through networks of tunnels; in container ports and construction zones the sky is filled with metal; buildings cast GPS shadows just like they cast sun shadows.
Rather than a single global coordinate frame that covers the entire earth, navigation in the built environment requires many small, short-range, but highly precise coordinate frames — all stitched together into an emergent network by software. Humatics calls this navigation microlocation and is bringing centimeter- and millimeter- accurate localization to the built environment. Microlocation will not replace GPS (and will use the best GPS data it can get), but rather fill in the places it does not work.
Microlocation fills the gaps in the built environment, where GPS does not work
Humatics microlocation works by embedding small, inexpensive beacons into the built environment, and enabling mobile devices like robots or mobile devices, to measure precise distances to those beacons. As in GPS, we measure the precise time of flight of small radio pulses, using covert, coded, ultra-wideband pulses. But where GPS measures those pulses over more than ten thousand miles of propagation (including through the complex atmosphere), we measure them over only a few tens or hundreds of meters. This shorter distance enables the signals to be thousands of times more precise than GPS and much more robust to multipath effects (where the signals bounce off of irrelevant objects), jamming, and other interference. In a typical installation, our “path” is 200,000 times shorter than GPS, and correspondingly more accurate and robust to multipath. Humatics microlocation combines radio pulses with inertial sensors and/or other data on the mobile or handheld into a software sensor fusion solution.
Most robots spend half their computing power, energy, and cost just figuring out where they are. Tell any robot where it is, and you’ve done half its job for it, freeing its machinery to concentrate on getting its interesting work done.
Now imagine a robot or a handheld that knows where it is within a factory, a warehouse, or a city down to the centimeter. Or even down to the millimeter. When robots know where they are, they can coordinate with people in ever more graceful, efficient ways. When people know where they are, they can navigate complex environments quickly and precisely. When subway trains know where they are, they can dramatically increase their throughput of people while maintaining high levels of safety.
At Humatics, we’re hard at work at several “hair on fire” uses cases for industry and mobility today. But we’re also aware that as microlocation proliferates robust and precise navigation to the built environment, it will change the world and how we experience it in ways we can still only imagine.