LiFi - The online light bulb

27 September 2016

Interview with

Professor Harald Haas, The University of Edinburgh

Wearable sensors and the internet of things are vastly increasing our ability to pureLiFicollect and analyse information. But with more and more connected technology, we could be heading for a problem. That problem is bandwidth - we're in danger of running out of radio frequencies to transmit all of our data. Earlier this week, Chris Smith talked to Harald Haas at the University of Edinburgh, who is proposing a new solution. It's called LiFi,  and it uses - you guessed it - visible light...

Harald - The traffic through our communication networks has increased rapidly so that it will take a normal human being five million years to watch all the video traffic that is sent through the networks in one month. We have an exponential growth in data that we transmit through our networks. Many businesses are centred around data and access to data and this is sort of the lifeline of the future economy.

Chris - Yes, indeed! Because in an article you've written and I quote, you've said "by 2020 the data deluge is predicted to increase to forty four zetabytes" -  I've never even heard of a zetabyte. You say "nearly as many bytes as there are stars in the universe."  So basically, we've got a data traffic problem; everything is producing data and we haven't got pipelines big enough to put it all down.

Harald - Yes, that is absolutely correct! That's why we at the University of Edinburgh are looking at different ways of transmitting data, i.e. looking at different pipes than we know at the moment. The classic pipe at the moment is the radio frequency spectrum but the radio spectrum is part of a larger family, and the family is the electromagnetic spectrum. And a big fraction of the electromagnetic spectrum is also the visible light spectrum which we have been exploring as a resource for data transmission.

Chris - So you're saying that rather than send things via WiFi radio, we could use the light we can see?

Harald - Yes, that is absolutely true. We use the visible light, the lights that are around us - the lights in our ceilings, the street lights, and these particular lights are becoming more and more LED based lights. And LEDs are electronic devices and, therefore, they can be used to provide illumination but this feature of - it's an electronic device also allows us to change the intensity much, much faster than we can even think or see, and these very high speeds at which we change the intensity, allows us to transmit and encode data onto the light that illuminates us.

Chris - So, effectively, the light in the ceiling could be made to get a tiny bit brighter or a tiny bit dimmer very, very quickly and that would carry a code which would transmit information to say my laptop?

Harald - That's absolutely correct.  And that laptop could have integrated another light source that is outside the visible range in the infrared spectrum and transmitting data back to the ceiling using the infrared, then we a bidirectional link.

Chris - And if we use ceiling lights in this way, how much data can we get through them - how fast can they transmit information?

Harald - That significantly depends on what kind of light sources we use. The fastest speeds that we can get, at the moment, is ten gigabyte per second from a single colour. So if we have red, green, blue creating white light, you could think of thirty gigabyte per second.

Chris - And the way you would see this being deployed is that say, a company, an office, public space, would have these light sources. They would be themselves connected to the internet and they would be looking at what light is coming out of people's computers in the room which would be people sending data to that hub, if you like. And it would then make the request onto the internet, the data would come back off the internet, and then the device in the ceiling would make the light flicker in just the right way to send the data down to your computer?

Harald - Yes, this is correct. Although lights in the ceiling would not be connected with mains power, they would be connected with a data cable - it's called power over ethernet so the data connection provides the energy to illuminate. So we have a data connection to the light and the data connection to the light is connected to the internet. So we have basically the internet brought to the light, and the light itself is then converted into flickering light and that is then captured by the smartphone and you get your data from the website request.

Chris - Now what about if you've got that really rubbish desk in the office that one-wants, the one that's in full glare of the Sun or the one in the corner that's all shady and dark - will it still work?

Harald - It's very important to recognise that sunlight will only very, very marginally affect the data rate, so that's the first point. And the second point is the light flicker is very, very small compared to the entire light output. But that then also means that we can dim the light down to very low levels while we still maintain the capability of transmitting data at very high speeds.

Chris - That was going be my question in terms of it doesn't sound good for the environment if we have to run all the lights all the time during the day but if we can turn the intensity of those light right down so it doesn't waste a lot of energy but we still get back the ability to transmit the data, then it's a good idea.

Harald - Yes, that's absolutely the case. We can reduce the intensity and then that's why we can also use it in street lights outdoors.

Chris - Now this sounds tremendous but have you got any evidence that outside of your laboratory, this will work?

Harald - Yes. There's been a press release, in fact, this week. We are equipping a large office of a security firm in Paris. They don't use WiFi because WiFi penetrates through walls, whereas light will not go through a wall and, therefore, it's much more secure to use light for data communication.

Chris - So, it took about ten to fifteen years for WiFi to become ubiquitous, so are we looking at the same sort of trajectory for LiFi?

Harald - Yes, I think it will be faster in my view. It took a hundred years for the landline telephone to have ubiquitous coverage, it took fifteen years for WiFi, and I predict it takes about five to eight years for a full uptake of LiFi.

Add a comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.