Systems and Networking

Nsf Grant to Help Point Way to 5g Wireless


Network providers and users have just gotten fourth-generation (4G) wireless under their belts, and already 5G wireless is on the horizon. It comes not a moment too soon as demand explodes for wireless devices such as smart phones and bandwidth-hungry applications. Experts say that by the end of the decade the disjointed and narrow bands of spectrum now devoted to wireless will be saturated and further growth will be impossible. Cisco Systems predicts that 25 billion wireless devices will be connected by 2015 and 50 billion by 2020, an average of more than six per person worldwide.

Exactly how 5G will work is not yet known, but researchers at the Polytechnic Institute of New York University (NYU-Poly) and Auburn University aim to find out. The U.S. National Science Foundation (NSF) in October awarded $500,000 to two researchers at the universities to "gain a deep understanding" of radio communication in the 60-GHz band. The three-year project will augment other research at NYU-Poly in the broad ultra-high frequency (millimeter wave) radio spectrum that runs from 28 GHz to 220 GHz. Millimeter-wave bands have the potential for increasing the spectrum available for wireless by several orders of magnitude.

Millimeter-wave wireless, which would operate at wavelengths from tens to hundreds of times shorter than those used in conventional cellular communications, has a number of advantages, says Theodore (Ted) Rappaport, director of NYU Wireless, a research center at NYU that has received other grants from NSF’s Networking Technology and Systems program. Opening up a much broader swath of spectrum, along with highly directional "beamforming" antennas that the shorter wavelengths permit, will cut infrastructure costs, prolong battery life, lower the probability of outages, and support much higher bit rates by many more users in any given area. Copper and fiber network backbones will give way to wireless, leading to a convergence of Wi-Fi and cellular, Rappaport adds. "This is an entirely new field," he says. "It’s going to be very big."

One of the key enabling technologies that Rappaport and his colleagues will explore will be tiny, high-gain antennas. At 60 GHz the antennas are the size of a fingernail and at 200 GHz the size of a freckle, Rappaport says. A portable device might then contain "many dozens" of antennas, he says. The antennas can be built in multiple arrays on a silicon CMOS chip and made highly directional and adaptive, so that narrowly aimed beams can be steered in just the direction needed. "Spatially locating where the signal is coming from, and combining the strongest signals from different paths, are really key," Rappaport says.

Those properties make the antennas good choices for applications such as collision-avoidance systems where a car must know and respond to the motions, directions, and distances of other vehicles. Similarly, in a cellular application, a device can find another device that is able to route a call around an obstacle, then communicate with it over a narrow beam that does not interfere with nearby signals. Developing the algorithms and protocols to do that will be the focus of project collaborator Shiwen Mao, a professor of electrical and computer engineering at Auburn.

Mao will take the models and designs that Rappaport builds for physical components such as antennas and figure out how to link them together to form multi-node networks. He says he will address questions such as, "How do you schedule, given a network of devices, which ones should transmit and which ones listen, and how do you reduce overhead to get the most efficient scheduling? Do they form one-hop networks or multi-hop mesh networks? How do they form a good topology and maximize network capacity?"

At a higher level, Mao will work on applications, such as uncompressed high-definition video streaming, enabled by the huge new spectrum capacity. "This is an unexplored area," he says. "The benefit is you can get rid of HDMI cables, and you can simplify by not having video decoding at the display device."

Attenuation of wireless signals by atmospheric oxygen and water limits propagation at various places in the millimeter-wave spectrum, especially at the 60-GHz band that is the focus of the NSF award. That makes it suitable for relatively short outdoor distances up to 100 meters or so, and for indoor applications. However, in August the Federal Communications Commission allowed for the use of higher antenna gains, allowing for great outdoor distances. But the 60-GHz band is still free of FCC licensing requirements.

These technical and regulatory properties of 60 GHz will enable important future applications, Rappaport predicts. For example, he speaks of "information showers" in which short-distance transmitters could beam safety and navigation information and advertising to passersby from every lamp post in Manhattan.

Indoors, Rappaport says that wireless connections among computer components, such as disk, memory, and CPU, will spring up at various scales. "There’s no reason you have to bring a whole laptop somewhere to bring memory," he says. "Why can’t your memory be in a credit card-sized RFID-type device that connects to whatever computer happens to be near?" Similarly, he says, wireless could replace miles of costly fiber and copper wire in a big data center with less cost and power consumption and more flexibility. That would also enable the isolation of heat-intensive components for more efficient cooling.

Although 60 GHz is not suitable for communication over long distances, Rappaport says that the industry for years assumed that no part of the millimeter-wave spectrum was usable for outdoor cellular telephony because of signal attenuation by oxygen and rain, and that millimeter-wave signals would not propagate around buildings. But no one until recently tested those "myths," he says. "In fact, we have known it’s not at all the case. At 28 GHz, 38 GHz, and 72 GHz, there are huge swaths of spectrum where there is virtually no atmospheric attenuation. Urban areas are surprisingly suitable for high-bandwidth communications when directional antennas are used."

Rappaport says that the much greater bandwidths available at millimeter-wave frequencies will allow instantaneous data rates in the 5-Gbit to 10-Gbit per second range, and possibly even greater, in comparison to today’s 4G LTE cellphone systems that offer tens of Mbit/s in the best case. "Eventually," he says, "I see millimeter-wave spectrum bands carrying data rates as high as 20 Gbits per second to 100 Gbits per second."

In addition to funding from NSF, the NYU Wireless program has been funded by a number of sponsoring companies, including Samsung Electronics, which earlier this year introduced what it said was the world’s first adaptive array transceiver operating in the millimeter-wave bands for cellular communications. Using 64 antenna elements, the 5G device transmits at 28 GHz at up to 1.056 Gbit/s to distances up to 2 kilometers, Samsung says.

Work by Rappaport "provides strong evidence of the feasibility of high-frequency bands for outdoor cellular applications," says Wonil Roh, director of advanced communications for Samsung. "The largely under-utilized vast amount of spectrum in the millimeter-wave bands can be the most effective answer to the problem of ever-increasing mobile data traffic."

Gary Anthes is a technology writer and editor based in Arlington, Va.  

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More