Let there be Li-Fi: managing high-speed internet above radio frequencies
Image credit: Dreamstime
If you have LED lights you are on the way to getting faster internet, although switching to light from radio for communications may be a bigger problem than it looks.
Radio frequency communication has a problem: it’s running out of headroom. Traditionally, its most valuable property when delivering data from one user to another is you do not have to aim the device in the right direction. Mobile users can move around as much as they want and they will still receive signals from a basestation just as long as they are in range. But that comes at the cost of increasing the interference when more than one user wants to use the same part of the RF spectrum.
Researchers have come up with increasingly sophisticated ways to code data so that it makes the most of each available RF band but they have butted up against the theoretical limit. To deal with the problem, designers of cellular communications decided to reduce the size of each cell and so cut the amount of interference that eats into the available bandwidth even though it is more expensive to achieve: the number of basestations you need spirals upward.
Another way to get more data over the airwaves is to move up in frequency. Telecom operators plan to make use of millimetre-wave transmissions, where the signals are much more directional: active antennas are needed to steer them, which tends to cut mutual interference. The bands are also able to squeeze in much more data than those in the conventional sub-3GHz space. Although the military snaffled the best slots in the mm-wave region, there is a lot more available spectrum between 10GHz and 100GHz. But go even further up the frequency range, into the infrared and visible-light range, even higher data rates are available and you do not have to ask anyone to use it.
Cone of protection
When the IEEE started work on what would go into a standard for Li-Fi communications, the task force looked at reasons why people might swap from RF to light. One is security.
Professor Harald Haas of the University of Edinburgh, a keen proponent of light-based communication, maintains the technology has a high degree of innate security for a couple of key reasons. One is that light is easy to block: if you are on the wrong side of a wall, you will have no way to pick up on transmissions. Wi-Fi signals, on the other hand, readily leak through many solid objects although the higher-frequency bands are easier to block.
A further reason for Haas’s confidence in the higher security of Li-Fi is that even if you are in the room, interception is still tricky to pull off. At a seminar organised by the Universidad Internacional Menéndez Pelayo (UIMP) in Santander, northern Spain, in late June, Haas argued that practical interception of encrypted Wi-Fi transmissions demands the attacker not just be able to see the downstream transmissions from a Li-Fi-enabled light bulb, but that of the return transmitter aimed at it. “The eavesdropper will need to be very close to you,” he argued.
The light cone itself could be very small. One of the IEEE’s proposed use-cases at home is a Li-Fi desk lamp with the user’s computer sitting directly underneath. However, as this requires a wired connection to the lamp, there remains the question of why not take that cable all the way to the device itself?
The idea of using light is nothing new, though not at the data rates Professor Harald Haas of the University of Edinburgh proposed in 2011 when he coined the name Li-Fi as a nod towards the still-dominant Wi-Fi RF networking standard. Haas alludes to an invention of Alexander Graham Bell’s that preceded the telephone, the photophone, which encoded the vibrations of a speaker’s voice onto light using a mirror. Bell successfully demonstrated that it could work over distances up to a kilometre.
Aside from its limited data rate, Li-Fi is quite similar to the TV remote control, the most pervasive example of digital light-based communication we have today. Like the remote control, Li-Fi does not need an unbroken line of sight – light bouncing off a wall can still work, though this might lead to more errors in the signal picked up by the receiver. As a result, Li-Fi systems will need to trade off error correction against speed. If they do not have a direct line of sight and suffer from flashes coming from other light sources that impinge on the detector, the effective data rate will drop.
Visible-light or infrared wireless links have also been used to take care of high-speed communications across wide rivers and between tall buildings scattered across a campus. The IEEE 802.11 wireless-networking standards committee, which looks after the Wi-Fi standard, has decided to include this kind of wireless backhaul as a use-case for the Li-Fi standardisation effort it kicked off earlier this summer.
Among the systems looking to exploit this use of light-based communication is one developed by Fraunhofer’s Heinrich Hertz Institute (HHI) together with Japanese telecom company Sangikyo. They designed the system to be able to support communications across a shopfloor or through subways without laying cables, with a peak data rate of 750Mbit/s to provide a balance between speed and robustness.
As with Wi-Fi, the IEEE 802.11bb task force has volume markets in the home and offices as its main focus. It proposes a system that, at minimum, supports data rates of 10Mbit/s but, through advanced coding schemes analogous to those used in the latest RF modems, scale to a peak of 5Gbit/s: almost ten times faster than the current maximum on Wi-Fi. In many cases, Li-Fi would operate alongside Wi-Fi, acting not just as a faster channel for anyone in sight of a light that can support it but a more secure one. The return channel would also be light-based but use a smaller emitter operating in the infrared region so that it does not interfere with the downstream signal or distract the user.
Volker Jungnickel, head of the metro, access and inhouse systems group at Fraunhofer HHI, says: “Standardisation is a prerequisite required to scale [Li-Fi] technology to larger volumes. However, behind successful standards such as Wi-Fi and 3GPP there is usually, in addition, a mass market and an ecosystem. Thus, besides standardisation we have to establish the market for Li-Fi and the ecosystem likewise.”
Finding applications for mass-market Li-Fi may be trickier than it looks even though Haas sees a willing ally in the lighting industry. That industry is faced with a shrinking market because the renewal rate for LED-based lamps is much lower than the difference in price they can charge for each unit. Light bulbs that can double up as routers for data lets them potentially increase the value of their product. And the core lighting technology barely needs to change at all. LEDs respond very rapidly to changes in current they receive from the electronic controllers in the bulb’s circuit board. All that is needed is a more complex controller that balances the demands of light generation and encoding data. But there are other obstacles.
One early concern was that of flicker. In practice the changes in light intensity are so small and fast, the human eye does not appear to notice them. However, Joël Thomé, general manager and senior research and innovation consultant at PISEO, points out that modulating the light to send data tends to shift the colour balance of the lighting slightly.
A bigger problem lies in getting data to and from the light fitting. Few homes are cabled for Ethernet and, where they are, the sockets are at plug height. Homeowners would need to run Ethernet cables into their ceilings to get to the Li-Fi bulbs. Would they want to incur the cost and inconvenience, or stick with Wi-Fi? Thomé says Li-Fi needs to show a strong advantage over existing RF communication systems to be deployed in the home: “For homes, Wi-Fi currently does the job very well. Why should we add the burden of additional cables? One might think about wireless communication between the LAN connection and the bulbs. But again, what is the competitive advantage over Wi-Fi?
“At the moment we do not really see any good reason why Li-Fi should be massively integrated in the consumer lighting market, especially in the short term,” Thomé adds.
The IEEE now has four groups working on different aspects of optical communication, only one of which is Li-Fi. Two focus on point-to-point visible-light communications that may be useful for telecom backhaul and low-speed transactions between internet-of-things devices. The fourth is optical camera communication (OCC), which treats pixellated light sources such as screens and LED lighting arrays, including those in vehicle headlamps, as active barcodes.
A human may just see a picture on an active billboard. But subtle changes in lighting may signal to a smartphone camera aimed at it other information, such as special offers or where to buy the product.
There are technical issues with the OCC concept. It takes more compute power to identify the transmitter than it does for the more straightforward problem of decoding the changes in light intensity that take place using Li-Fi. But OCC avoids the problem of having to fit the system with additional circuitry to decode light-based signals. When it comes to smartphones, the usage model is one already exploited for static QR codes and augmented-reality apps.
The major application for OCC may be in automotive. Cars today are routinely armed with front- and rear-facing cameras and use sophisticated image processing to pull important features out of the field of view. And they need to be powered-up while the car is moving. The additional overhead of homing-in on traffic lights, headlights or brake lights to see if they are also OCC transmitters is unlikely to be high, although brake lights introduce a complication. OCC can potentially use low-level signals so that lights that appear to be off are still useful for communication – but detecting them in the field of view may be trickier.
Proponents of OCC see those lights as being conduits for information on traffic flows and near-term intentions such as “traffic light will change to green in three seconds”.
The office environment possibly provides more fertile ground for Li-Fi. Lighting suppliers are already looking to building maintenance as a way of helping to keep them in business. Rather than sell lights and maintenance, the lighting supplier can provide a set of services delivered through the light sockets. A number of lighting specialists have already started to work on incorporating Bluetooth into their products so that the lighting grid can act as a feeder network for IoT sensors. They avoid new wiring by taking advantage of mesh networking in Bluetooth version 5.0.
Mesh networking lets Bluetooth messages hop from node to node until they reach their destination, possibly across an entire building. As a result, it needs no new wiring. But building owners might choose to take advantage of ability to use Ethernet cables to transfer power as well as data. The Power over Ethernet (PoE) standard was developed as a way of providing electricity to network devices so that they do not need a separate mains connection. Most offices already have cabling running close to the lighting grid that could support PoE and, with it, data delivering to Li-Fi-enabled bulbs.
Bluetooth overlaps with potential Li-Fi services in other ways. The RF standard, for example, has the concept of the beacon: a device that provides location services, such as the ability to tell devices in range where they are and what services are available to them. Li-Fi proponents want to do the same.
Bought by Philips in 2016, French specialist Luciom came up with the Li-Fi Tag – using a relatively low-speed code imprinted on the light to act as an indoor guide and asset tracker. One of the first applications Philips found for light-based communications was in a trial for the French supermarket chain Carrefour. Codes picked up from the light fittings overhead by a smartphone’s camera would tell an app where the user is in a store and provide a map to where they would find items on offer or something they want to buy.
Although office and retail could provide a sizeable market for Li-Fi in the future, the near-term applications may be in more specialised markets. Thomé points to healthcare and aviation as possible targets.
Last year, Airbus said it is experimenting with Li-Fi as it provides a way of delivering high-speed data to each passenger’s seat without having to install large quantities of cabling that will add to the aircraft’s weight. As each passenger has their own light above their seat, the passenger aircraft may be one of the environments best suited to Li-Fi communications.
In hospitals, a big concern that prevents widespread deployment of Wi-Fi or Bluetooth is that of electromagnetic interference affecting the more sensitive instruments. Explosion-sensitive environments on oil rigs and industrial plants that handle fine powders and volatile chemicals similarly cannot make use of high-frequency RF communications and also demand stringent protection around data cables. Light-based messages are much safer. One of the use-cases explored by the IEEE 802.11bb task force is for communicating with robots sent into gas pipelines – again, because of the lower risk of explosion.
Fraunhofer HHI has focused more on industrial applications than on consumer for the near term. Jungnickel says it is hard to predict where Li-Fi will succeed. “We just believe that the original idea, to jump in one step from zero market into a mass market for consumer applications, is too high. This requires from the beginning low cost, low power-consumption and high volumes. That’s why we were seeking for other use-cases and found them in the industrial area. Our biggest challenge is to find real customers and work with them to introduce the technology into real markets.”
As well as working on its wireless-backhaul technology, Fraunhofer HHI put infrared Li-Fi transmitters and detectors into a manufacturing cell at car company BMW’s plant in Munich in a project called Owicells supported by the German government. As a robot carried out its usual activities – welding and test parts – and it moved around the 5 x 5m2 cell, it exchanged data with an array of infrared emitters and detectors scattered around the perimeter. The team used multiple transmitters and receivers to improve the reliability of the 100Mbit/s link, with most of the data being beamed from the robot to the cell as it welded parts together.
The Fraunhofer team found the bright flashes from the welding equipment did not affect reception and believe this was because the flashes are short compared to the time it takes to transmit each set of data symbols. The use of multiple emitters helped avoid breaks in communication as the robot’s arm moved, sometimes blocking the line of sight to the nearest detector.
The technology is moving ahead and will most likely be spread across the visible spectrum, with infrared being used in many cases because it is less distracting. However, the ability to use standard LED light fittings will avoid the need to budget as tightly for energy as with RF – because most of the power will be used for illumination. “In the end,” says Thomé, “one might have different scenarios for Li-Fi deployment depending on what drivers will appear in the future.
“The technology is available and is not an issue for widespread adoption of Li-Fi,” he adds. But he notes that standards and the demonstration of competitive advantages over wireless technologies that are already in place are the challenges Li-Fi needs to overcome. But, in the long-term, Li-Fi may become a necessity simply because RF has run out of space and Li-Fi and optical fibres have to come together to keep pace with the data demand.
Space is the place
Li-Fi could reach into deep space. In 2013, Nasa launched a satellite destined for lunar orbit that carried a novel way of relaying data back to the Earth. The Lunar Laser Communication Demonstration used an infrared laser working at a similar wavelength to those used in long-distance fibre-optic communication to send data back to Earth.
The 622Mbit/s link coped with clouds by switching between ground stations. The team found it was able to switch and resynchronise with each new station without having to fall back on a radio side-channel, which helps simplify the system. For space missions a shift from RF to infrared or visible light communications makes possible a hundred-fold increase in bandwidth.
The next step in the Nasa project is the Laser Communications Relay Demonstration project, which will be used to explore the feasibility of having satellites in orbit capture signals from deep-space probes and convey them to ground stations.
Visible light may prove the key to making it possible for robots operating underwater to talk to each other more easily. RF scatters so readily in water that it is hard to establish reliable communications except at very low frequencies – with equally low data rates. Water also absorbs visible light but, using Li-Fi-type modulation, blue-green lasers can send data much further than all but the highest energy electromagnetic waves. Even with blue-green lasers, scattering remains an issue but Li-Fi could be used for tens or hundreds of megabits per second communications over distances of tens of metres.