We use many wireless devices indoors—for better or for worse. Cellular telephones replace landlines in many homes; public Wi-Fi “hot spots” slow down during times of peak demand; laptop users wonder whether the neighbors are picking up their radio-frequency (rf) signals.

The next generation of electronic devices, however, may connect to the Internet and the telephone network through a different region of the electromagnetic spectrum: light. The growing acceptance of indoor solid-state lighting provides a potential platform for a new means of delivering data to consumers.

Because the devices would operate indoors, engineers and users would not need to worry about signal attenuation due to fog, air pollution and other atmospheric hindrances that can affect free-space optical systems between urban buildings or on the battlefield. Since light cannot pass through walls the way radio signals can, light signals could be more resistant to poaching than Wi-Fi emissions. And no wavelength of visible light needs to be licensed from a governmental communications authority.

Researchers in the United States, the United Kingdom, Germany, South Korea, Australia and other countries are pushing this “Li-Fi” light-based technology forward. International standards bodies are rushing to keep up with the development of visible light communications (VLC), also known as optical wireless communications (OWC).

The first commercial applications may involve indoor positioning: showing workers the positions of mobile equipment in a large building complex, perhaps, or offering shoppers special deals based on their location within a store. Outdoors, LED headlights and tail lights on automobiles could allow the vehicles to “talk” to each other and potentially reduce the chance of collision.


figureMobile data explosion: Increase in worldwide mobile data traffic, projected to 2016, shown in exabytes per month (1 exabyte = 1 million terabytes). CAGR = compound annual growth rate.

Wi-Fi is not going away, however. Li-Fi developers say that they envision asymmetric, hybrid wireless communications systems combining LED hot spots with traditional rf routers. Still, by allowing delivery of high-bandwidth data such as video via light, the technology of VLC or OWC may take some of the burden off the ever-more-crowded rf spectrum and add new capabilities and value to solid-state lighting installations.

A crowded rf spectrum

In news articles covering mobile wireless technology, phrases like “spectrum crunch” are popping up more frequently. The proliferation of wireless devices that use rf signals has prompted telecommunications carriers to seek ever more capacity.

According to Thomas Little, an electrical and computer engineering professor at Boston University (U.S.A.), people are carrying an ever-increasing number of devices that consume rich media, including high-definition video, in ways never before possible. “It’s at the point where we inhibit the ability of the devices to continue to evolve,” he says.

The Cisco Visual Networking Index, which tracks global data traffic, found that global mobile data traffic grew 70 percent during 2012, the last year for which information is available. In the same period, average smartphone data jumped 81 percent, and for the first time, videos accounted for more than half of mobile data traffic. In 2012, mobile device users were consuming 885 petabytes per month, but that monthly total is predicted to soar to 11 exabytes by 2017.

The U.S. National Telecommunications and Information Administration’s official frequency allocation chart, which runs from 3 kHz to 300 GHz, shows that the radio spectrum is already carved up, except for a few unallocated frequencies at the extreme high and low ends. Television and radio broadcasters, public safety and military communications systems, satellite and marine navigation systems and radio astronomers all need to operate at precise frequencies without interference.


figureHarald Haas experiments with direct-current optical (DCO) OFDM. The OFDM can be seen on the oscilloscope. According to Haas, DCO-OFDM effectively removes intersymbol interference and multipath interference.

Mobile broadband carriers are always asking governmental regulators for more bandwidth—but the laws of physics don’t allow creation of additional radio frequencies. “Nobody will buy service from you if it keeps getting congested,” says Mohsen Kavehrad, an electrical engineering professor at Pennsylvania State University (U.S.A.) who studies optical communications.

Various studies have shown that consumer-oriented wireless devices tend to be used indoors—in homes, offices and, to a lesser extent, inside transportation vehicles. Those settings also need artificial lighting, and in recent years solid-state lighting has developed into an effective alternative to incandescent and fluorescent bulbs. The growth in affordable LEDs has set up a synergy between the need for illumination and the need for wireless data access points. Little, his colleagues and his competitors are jumping on that opportunity.

Li-Fi: An idea gets started

Sending information via light beams is not a new idea. The inventor of the telephone, Alexander Graham Bell, experimented with a “photophone” in the 1880s. Starting in the 1970s, free-space optical communications used lasers, usually at infrared wavelengths, to lessen noise from sunlight and to send signals between buildings or battlefield stations no more than 2 or 3 km apart. In the 1990s, Joseph M. Kahn and his colleagues at Stanford University (U.S.A.) did pioneering work in indoor wireless infrared transmission among other topics in free-space optics.

A decade ago, Harald Haas, an engineering professor at the University of Edinburgh (Scotland), was studying third-generation (3G) mobile communications systems when, he says, he started thinking about Li-Fi after a discussion of rf spectrum allocation at a World Radiocommunication Conference in Switzerland.

From his telecom work, Haas had learned of orthogonal frequency-division multiplexing, (OFDM), an encoding scheme used in rf but not ideal for it. Compared to other types of multiplexing, OFDM has a high peak-to-average power ratio (PAPR), but Haas thought this disadvantage could be turned into an advantage in visible light communications. Simultaneously, Jean Armstrong, an engineering professor at Monash University (Australia), embarked on pioneering studies of OFDM in optical communications, both within optical fibers and in free space.

Large room equipped with both Wi-Fi and VLC access points.

While still at Jacobs University, Bremen (Germany), Haas and his students experimented with visible light signals using OFDM and higher-order modulation techniques such as quadrature amplitude modulation (QAM). After moving to Edinburgh in 2007, he embarked on a proof-of-concept demonstration of high-speed video transmission with OFDM. At the 2011 TEDGlobal conference in Edinburgh, Haas gave his first public demonstration of the technology, using an off-the-shelf LED lamp and an inexpensive receiver to stream video. “That for us was a big joy and a big moment of success,” he says.

In the United States, several multi-institution partnerships are investigating OWC as part of their studies of energy-efficient “intelligent lighting.” Among them are the Smart Lighting Engineering Research Center, led by Rensselaer Polytechnic Institute with Boston University and the University of New Mexico. Penn State and Georgia Institute of Technology run the Center on Optical Wireless Applications, led by Kavehrad.

How Li-Fi works

Conceptually, a Li-Fi hot spot is almost as simple as a signal lantern. Data is piped, via a power line or Power over Ethernet connection, to an LED luminaire outfitted with signal-processing technology. The luminaire shines directional light onto users and their devices, streaming the data by varying its output intensity so rapidly that the human eye does not notice it. (The human eye cannot detect a light source’s on-off flickering beyond roughly 100 Hz, according to Little.) A photodetector on a user’s device reads the tiny fluctuations in light intensity and converts them into an electrical signal.

Li-Fi systems rely on the incoherent light generated by LEDs, which is a restriction not found in laser-based free-space optical systems. Laser signals vary by phase as well as amplitude, resulting in a complex-valued signal. However, with incoherent light, it’s only possible to change the amplitude and thus the intensity of the signal.

LEDs are semiconductor devices, so they react very rapidly to changes in the electrical current that powers them, for up to 20 MHz of bandwidth—four times the bandwidth of third-generation mobile carriers.

Existing cellphone and tablet cameras can handle low optical data rates, but to get the high data rates required by streaming-video and rich-media applications, portable devices will need photodetectors with additional signal processing and nonimaging lens designs. To build practical systems, engineers need to evaluate the degree of directionality of the transmitter and receiver and to specify whether there needs to be a line-of-sight path between the two. Hemispherical lenses may increase the field of view of the receivers on handheld devices.

Some of the OWC technical issues currently being discussed in journals include LED nonlinearities and the effect of pulse-width modulation dimmers. Haas says these issues do not necessarily result in greater complexity to the OWC system. His Edinburgh group has developed a variant of spatial modulation, called space shift keying, which he says results in a bit error ratio of less than 2 × 10–3.

Which LED?

VLC signals could be delivered by red-green-blue LED combinations, white LEDs or organic LEDs (OLEDs). Each has advantages and disadvantages, according to Zabih Ghassemlooy, professor of optical communications at the University of Northumbria, Newcastle (U.K.)


  • Well-known technology

  • Limited use due to difficulties in RGB balancing

  • Phasing out in lighting industry

  • High cost

  • Higher bandwidth

Blue chip + Phosphor

  • Popular for today’s general lighting industry

  • Standardized for illumination and communications

  • Low cost

  • Modulation can cause color shift


  • Emerging technology

  • Early stage of development

  • High potential

  • Very high cost

At a conference in 2012, Haas, Irina Stefan and Hany Elgala reported on an asymmetrically clipped optical OFDM indoor wireless system in an average-sized room under different overall brightness levels. Using continuous current reduction (CCR) instead of pulse-width modulation to dim the LEDs, the group realized a signal-to-noise ratio (SNR) of 9 dB on a desktop receiver located directly below one of the room’s luminaires.

“In a room where you see objects,” Haas says, “then you can also communicate.”

Potential advantages of Li-Fi

One of the advantages of Li-Fi over rf wireless is as obvious as the nearest wall. Since visible light does not pass through opaque objects, two people using the Internet over Li-Fi in adjacent rooms cannot detect each other’s signal. “If I open up my Wi-Fi, I see at least five neighbors’ Wi-Fi signals,” Haas adds.

As Little explains it, rf and VLC wireless systems take different approaches to spectrum reuse. Each user of a Wi-Fi hot spot competes to use the same batch of radio spectrum as his or her neighbors, because of their spatial overlap. The more users on the hot spot, the smaller a “slice” of bandwidth each one gets.

Since Li-Fi users in separate rooms don’t interfere with each other, they can reuse the bandwidth more often than in a radio frequency system with signals propagating through walls, according to Haas. The signals would also be less vulnerable to people trying to break into the system. Existing data encryption protocols would work with either radio or optical wireless signals.

The upstream direction

VLC will not make rf as obsolete, however. “If your phone is in your pocket, you can’t get light-based communication,” Little says. The mobile network of the near future may use both the radio and visible spectra in complementary ways, and devices will detect the presence or absence of both kinds of signals.

Another important open question is how the upstream direction is handled. Wi-Fi networks are bidirectional—but while Li-Fi might allow for streaming of data to mobile devices, how will data get from those devices back to the network? Little believes that the most promising deployment scenario uses available rf channels as the backchannel to a local rf (Wi-Fi) access point. “This is consistent with the active approach to ‘converged heterogeneous networks’ in which multiple media are used opportunistically to get the best performance,” he says.

A light-based backchannel may be possible, according to Little, and some Li-Fi researchers are experimenting with this symmetric model. However, a visible-light backchannel may fail due to glare and other issues. Some researchers are experimenting with an infrared-based Li-Fi backchannel.

Haas envisions future all-optical communications systems in which indoor spaces contain few if any rf base stations, but are filled with LED fixtures that are switched on continuously, delivering both light and gigabit-class data. Others say that the visible-light transmissions will serve, at best, as a unidirectional drop for video consumption, but—since devices don’t come equipped with LED lights of their own—VLC will never break free of the need for rf uplinks.

Potential markets

Late last year, a British research team tested a single-LED wireless OFDM-based VLC link for speed. Using a 50-μm-diameter gallium nitride LED emitting at 450 nm, the group sent the encoded light signal over a distance of 5 cm and processed the received signal on a desktop computer via MATLAB. The researchers realized a data rate of 3 Gbps, but acknowledged that the tiny power output of the micro-LED restricted the distance over which the link could operate. The project involved not just Haas and his Edinburgh colleagues, but researchers from the universities of Oxford, Cambridge and Strathclyde. The team reported on their proof-of-concept experiment in IEEE Photonics Technology Letters.


figureThomas Little, of the Boston University College of Engineering, says that some Li-Fi products under development are not any more complicated than typical Wi-Fi implementations.

Several Li-Fi researchers have founded companies to turn the technology into convenient products, which are still in the prototype stages. ByteLight, founded by some of Little’s former Boston University students, is developing a system to send messages about hot bargains to shoppers’ existing smartphone cameras based on their locations—a kind of indoor “light positioning system.”

Haas’s startup company, which recently changed its name from pureVLC to pureLiFi, has an initial product offering that promises 5 Mbps data rates for both infrared uplinks and visible-light downlinks, thanks to p-i-n diodes in dongle-type receivers that can be attached to existing portable devices. In general, p-i-n diodes are cheaper and less sensitive to temperature fluctuations than avalanche photodiodes.

Another potential market for VLC is the healthcare industry, since hospitals have long banned cellphones and Wi-Fi hotspots that could interfere with sensitive medical equipment. Li-Fi could provide communications to patients and visitors, while lower-data-rate light-based systems could track positions of moveable equipment from wheelchairs and gurneys to emergency-room “crash carts.”

New airliners, Kavehrad notes, will be equipped with LED lighting, so the most important part of the VLC infrastructure will be already in place. The solid-state luminaires could give airlines a way to provide telecom data streams without radio-frequency signals. The small spotlights above each jetliner seat would act as individual, short-range Li-Fi hot spots.

In museums and galleries, short-range Li-Fi signals could give viewers information about each nearby work of art. Underwater, divers could communicate with each other through their LED headlamps, and mobile robots navigating a factory floor could get position data, probably to a resolution of a few centimeters, from the overhead light fixtures.

Moving toward commercialization?

Haas wants to make VLC part of the international standard for fifth-generation (5G) mobile communications, which may hit the market by the early 2020s. He has organized discussions on 5G at several IEEE conferences and would like to get more optical scientists involved in the standardization effort.

Little says that some of the Li-Fi products under development are not any more complicated than typical Wi-Fi implementations and use similar signal-processing and protocol stacks. Although today’s prototypes are indeed complex and expensive, he adds, the path to low cost comes from production scaling, which will ultimately happen outside of the research lab.

The building blocks for VLC and hybrid optical-radio networks—LED luminaires, power-line communications, Power over Ethernet—already exist. Clearly, many issues in network interoperation, security and backwards compatibility with existing devices need to be addressed. However, Haas believes the growing radio “spectrum crunch” will eventually drive Li-Fi adoption. “It is now really an engineering task and not so much a research task anymore,” he says.

Patricia Daukantas is a freelance writer specializing in optics and photonics.

References and Resources

  • J.M. Kahn and J.R. Barry. “Wireless Infrared Communications,” Proc. IEEE 85, 265 (1997).

  • J. Armstrong. J. Lightwave Tech. 27, 189 (2009).

  • H. Elgala et al. IEEE Commun. Magazine 49(9), 56 (2011).

  • H. Haas. “Wireless data from every light bulb,” TEDGlobal 2011 presentation.

  • I. Stefan et al. “Study of Dimming and LED Nonlinearity for ACO-OFDM Based VLC Systems,” Proc. of Wireless Communications and Networking Conference (WCNC 2012), Paris, France, 1-4 April 2012.

  • H. Elgala and T.D.C. Little. Opt. Express 21, 24288 (2013).

  • T.Q. Wang et al. J. Lightwave Tech. 31, 1744 (2013).

  • J. Armstrong et al. IEEE Commun. Magazine 51(12), 68 (2013).

  • D. Tsonev et al. IEEE Photon. Tech. Lett., published online 2 January 2014 (doi: 10.1109/LPT.2013.2297621).