Overloaded wireless-data networks are so common in crowded urban areas that many of us take them for granted. Which helps explain why a startup like Artemis Technologies, which promises near-perfect cellular reception to smartphone users no matter how crowded their networks, has gotten some rapturous attention over the past month.
Artemis announced its approach, which it calls pCell (for "personal cell"), at the end of February. Despite the hype, its technology is far from proven, and has yet to be demonstrated outside of carefully controlled settings. Even if it does work as advertised, there's no guarantee that wireless giants like Verizon and AT&T will embrace it, since scarce bandwidth allows them to charge higher data prices.
But the promise of pCell is unquestionably attractive, and while the underlying technology is counterintuitive, it's well-founded in communication theory. Artemis, too, has a pedigree; it was founded by Steve Perlman, the inventor/entrepreneur behind OnLive, WebTV and Apple's video encoding technology QuickTime.
So while Artemis is clearly out on the frontier, no one should dismiss pCell out of hand. Here's what you need to know about this technology and what it could mean for the future of wireless data (and, possibly, more).
1. Wireless Data Is Bottlenecked By Interference
If your LTE data slows to a crawl even when you have good reception, you can blame what telecom researchers call "the cocktail party problem." When there are way too many people in a room, there are too many signals—and too much interference—for you to hear the person you're talking to.
The same goes for wireless, where your experience is usually affected by the thousands of other people nearby also trying to connect to the Web. All those radio signals zipping from cell tower to individual phones and back routinely get in each others' way, sometimes canceling one another out entirely.
To see how this works (in a vastly simplified way), consider this photo of a lovely New Zealand harbor, taken as two wakes are crossing one another. Think of those crossing waves as the data signals heading for your smartphone and that of someone down the block.
Now let's look at a closeup of where the waves meet (click for a larger version).
See how the water flattens where the peaks of one wave meet the troughs of another? It's a phenomenon physicists call destructive interference, and something very similar happens when radio waves of the same frequency cross paths. In other words, those are your smartphone "dead zones."
Now multiply this effect by thousands of signals pinging around in some dense cell sites, and it's clear that dead zones are everywhere. Communications engineers have a number of tricks to keep you from ending up with no signal at all, but they all have one thing in common: They limit you and your phone to a mere fraction of the total bandwidth your local cell tower puts out.
2. pCell Actually Uses Interference To Its Advantage
Three years ago, Artemis founder Perlman released a white paper describing a technology he called DIDO, for "distributed input distributed output." DIDO purportedly allows wireless users on a network to use the full data capacity of shared spectrum with several other users simultaneously.
"DIDO profoundly increases the data capacity of wireless spectrum, while increasing reliability and reducing the cost and complexity of wireless devices," Perlman wrote. "DIDO deployment is far less expensive than conventional commercial wireless deployment, despite having vastly higher capacity and performance, and is able to use consumer Internet infrastructure and indoor access points."
DIDO—now pCell—basically exploits the flip side of destructive interference. Let's go back to those New Zealand waves for a second.
Now look where the peaks of both waves meet. This is also interference, but here it creates a larger wave—one that, were these radio waves, would correspond to a stronger signal. Artemis' technology aims to exploit such "constructive" interference in order to deliver a fast, clear signal to mobile users.
3. pCell Involves Some Very Serious Computation
One of the most interesting claims Perlman makes for pCell is that it sidesteps Shannon’s Law, an information-theory principle that establishes a limit to data transmission speeds over a given channel. In wireless, the main consequence is that mobile devices served by wireless networks of the same frequency interfere with one another and thus "divide up" the available signal, limiting their data reception.
pCell, however, purports to fine-tune the pattern of radio signals in a way that allows each device to receives the data capacity of the full channel. As a result, DIDO claims to pull off something normal cell towers can't achieve: Even when more users join the network, the data rate for each user remains constant.
Effectively, this is because the pCell system is designed to create pockets of constructive interference around every mobile-device antenna. Each such pocket—Perlman describes them as about a centimeter in diameter—would act as an independent data channel, and thus wouldn't be limited by interference from other devices.
The way Artemis describes it, there's some very heavy lifting going on behind the scenes. In place of a single cell tower broadcasting to all users within range, pCell would establish a network of smaller antennas, each of which would broadcast a precoded signal computed by a data center that manages all communications in the area. Each precoded signal would basically be gibberish by itself—but where they overlap, the resulting constructive interference would yield a clear, high-speed data signal for every device in broadcast range.
Now, this is most definitely rocket science. The pCell system would have to generate those complicated precoded signals in real time, taking into account the actual data being transmitted to users, their location, their movements, and interference from solid objects like concrete walls.
It's an extremely complex problem, and probably one of the biggest reasons to be skeptical about Artemis' claims. Perlman, of course, says his company has solved it.
4. pCell Wouldn't Require New Phones, But Would Replace Cell Towers
One of pCell's biggest selling points is that it would work with all current 4G LTE-enabled devices. That, of course, would be a huge advantage, as it wouldn't require everyone to ditch their current iPhones and Androids just to get a decent signal.
But the company also envisions a new line of “pCell-native” devices, which the company says are “faster than LTE with fiber-class latency” thanks to a low degree of power consumption. If this promise rings true, it could have a significant implications for newer and more dynamic mobile technologies such as wearables.
Cellular infrastructure, however, is another story. Today's mobile devices are reliant on cell towers for their signals, but they're costly in several ways. The average cell tower costs $150,000 to build, and takes a toll in worker fatalities, too. In 2012, ProPublica estimated there were an average of 123.6 cell tower worker deaths per 100,000 workers—more than 10 times the death rate across the construction industry as a whole.
Instead of cell towers, Artemis offers “pWave radios,” which are small but stylish weatherproof devices that the company says can be easily installed on indoor or outdoor walls or ceilings. Here's how Perlman described them at the first public demonstration of the technology at Columbia University in February:
Rather than having one cell tower, you put a few of these small pWave radios around an area and that's it. The transmissions, rather than being one transmission, intersect, and when they intersect, they create a tiny personal cell around every mobile device that's around a centimeter in size.
5. There's Just One More Thing
pCell could have even greater implications for other kinds of wireless technology. According to this detailed explanation of pCell technology by San Francisco data scientist Imran Akbar, pCell might also allow transmission of wireless power to for devices like phones, tablets, TVs and even motor vehicles. Should that pan out, smart devices might never need plugging in again.
Artemis' approach, however, isn't exactly unique. In fact, several other companies have invented solutions based off similar technology, and the theories for this system—also known as "network MIMO," "cooperative MIMO" and "cooperative beamforming"—date back to the early 2000s.
Perlman says Artemis will begin deploying pCell by "late 2014." But even if pCell can scale up to urban environments in a cost-effective way, that doesn't mean pCell is guaranteed to succeed. The best technologies don't always win, and Artemis has yet to conduct any large-scale demonstrations that might sway skeptics.
Artemis will also have to convince carriers and the FCC that its technology is worth the billions of dollars it would require to build base station radios and massive cloud infrastructure. It could be a long haul.
Most images courtesy of Artemis Networks; New Zealand harbor and wave images by Flickr user brewbooks, CC 2.0