Dr Sithamparanathan Sabesan, Cambridge University
Chris - If you've ever walked out from a shop and set off an alarm – I'm sure you haven’t for anything other than accidental reasons – then you've probably triggered an RFID, that's Radio Frequency ID tag system. The problem with these tags is that although they can identify an individual piece of merchandise and are therefore very useful, they only work over very short distances and that limits their use.
But now, new research from Cambridge University has found a way to dramatically increase this range and therefore, the usefulness of these pieces of equipment. To tell us more, Sithamparanathan Sabesan who is from Cambridge University’s department of engineering where he’s a research fellow, is with us to tell us how it works. First of all, thanks for joining us. What do these RFID tags actually do? How do they do what they do?
Sithamparanathan - RFID stands for Radio Frequency Identification. These RFID tags do not have an internal power source, you have to power them up for them to receive signal and retransmit signal for example. And so, we have antennas. They transmit an electromagnetic signal for them to basically receive the signal. And these RFID tags then have a unique ID stored in the microchip.
Chris - So, the energy comes in from a radio signal from a transmitter. It interacts with the tag which gets its energy from that, powers up a little microchip which sends a unique radio signal back to a detector and that says, “Hi. Here I am” and identifies that item uniquely.
Sithamparanathan - Absolutely.
Chris - So that sounds terrific. What is the problem with these tags? Why are they an issue?
Sithamparanathan - The real challenge here is then, because they don't have a battery, it’s very difficult for us to read this tag over longer distance, due to the fact that the signal not only travels in one single direction, but also, travels in multiple directions.
Chris - Because of in reflections off of walls and the floor, and so on.
Sithamparanathan - Exactly. So the results of that, the direct signal gets cancelled out by the reflected signals.
Chris - So this is the signal coming in from the transmitter that you're beaming to the tags. It will bounce off all the surfaces and in certain places, you're going to cancel out the signal and get a dead spot.
Sithamparanathan - Exactly and we call them dead spot. The reality is that, over a large area, you’ll have so many dead spot that typically, you can only read 60% of the tags in a conventional approach. And so, we developed a technology called dithering. We have multiple transmitters, they synchronise in a way that we are able to move this dead spot around within that area. As a result of that, I shall be able to read this tag with 100% accuracy. I won't miss any single tag over that large area.
Chris - Wow! So, where previously with a static array of transmitters, you would end up with a pattern of dead spots. It’s a bit like if you walk around your house with a mobile phone, it works in some rooms and not others. You’ve got a pattern of dead spots and so, if the tag is in one of those, you're not going to spot it. But with your system, you're moving the dead spots around in the environment, dithering them. So, even if a tag is in a dead spot one moment, the next moment it's not and you're going to get the signal back. And so, the trick is how you move the dead spots then?
Sithamparanathan - So we use multiple transmitters in a synchronised way by doing some clever advance signal processing at the backend and also synchronise them. These are able to move these dead spot very quickly because you got to move them very fast. It’s typically in milliseconds kind of speed. By doing so, over the 1 second period, I shall able to read all the single tags.
Chris - And this means that then, rather than just having to put in massively powerful transmitters which still wouldn’t get around the dead spot problem, you've now got a simple array, you use these very cheap tags and you can detect them over, I would say, probably much greater distances as well then.
Sithamparanathan - Yes. So far, we have demonstrated over 20-meter x 20-meter area with 100% accuracy. You could scale that to a large building, typically hundreds and hundreds of meter large buildings, you could have a small number of antennas and read all the tags over that area.
Chris - How do you actually read where the tags are? In order to scale it over a very big area like that, you're detecting a tiny signal coming out from a tag which could be anywhere in that environment? So is that a sort of triangulation trick then?
Sithamparanathan - Exactly. So, in order to locate the tags, that's another challenge. Reading a tag is one challenge we just discussed and locating the tags is another challenge. You could have triangulation. Even then, you're still going to have the multipath and dead spot problem, but we have another invention which enable us to triangulate these tags in a novel way. The results of that is that you can locate these tags within 1 meter location accuracy.
Chris - So supermarkets are going to be delighted because I'm thinking, all I need to do is I go into the supermarket, swipe my credit card and identify myself electronically to the supermarket with a sort of trolley or something and then every time I take something off the shelf, the supermarket knows where the object is because you can have these tags on every piece of stock. So a.) It’s good for stock control, b.) It stops shoplifting, and c.) It means that I don't have to queue up and check out. I just walk out.
Sithamparanathan - Exactly, so supermarket automatic checkout is one of the long term visionary application we’re focusing on. As you said, in the future, you don't need to stand in a queue. You just walk with your trolley and pass through that reader and it will give you the bill straight away, there’s no queue whatsoever. But there are other applications we’re currently launching, one of which is document tracking. Working with local councils where they want to see where every single file is within large office area...
Chris - I think also in a hospital because tracking down a patient’s notes – it sounds trivial but if you're in the operating theatre and the notes are on the other side of the hospital, knowing where they are to send someone to get them could be really helpful.
Sithamparanathan - I'm glad that you mentioned hospitals because that's one of the customers we’ve been speaking with very recently. They will also want to track their patients because they also want to learn how their diseases spread within hospitals.
Chris - Have you got any trial data to show that this will actually work though? It’s one thing to say, “Right, we think we can do this in supermarkets, we think we can make this work in a hospital to track people and notes.” What about the reality though? Have you got data on this?
Sithamparanathan - So very recently, we have developed and demonstrated for the first time a document tracking system. We were able to show that we were reading hundreds of files over that large area. We were demonstrating over 20-meter x 20-meter area. There's also another big launch that happened in London last week, for tracking containers and cars in over a large area and so, so far, we have a prototype system which is able to show that we have 100% coverage over that large area for the first time.
Chris - Amazing stuff, thank you very much. That’s Sithamparanathan Sabesan, from the Department of Engineering at Cambridge University.
This is a good idea. It is using spacial and temporal diversity to avoid multipath cancellation of a signal. I think the expansion of this technology into reading every item in a shopping trolley is key - this is the holy grail solution. I am not sure that operation over a wide area is so important although it could then work with fewer (or even just one) transmitter array which could be directed to one small region at a time to see which trolley has which goods. I guess, alternantively, all trolleys could be transmitted too but the receive range could be kept low with a receiver near each trolley as it passed by. graham.d, Wed, 25th Apr 2012
Raghusesha - Dr Chris normally mentions any recent articles which might have prompted the interview - why don't you try Google Scholar or Arxiv on Dr Sithamparanathan Sabesan imatfaal, Thu, 3rd May 2012
Most interesting extension of a passive device.
Absolutely right Clifford. You can see the usefulness of such systems and the potential demand for it but, as with many advances, there are always consequences that need to be considered. An example was fingerprint recognition that would make it hard for anyone to benefit from mugging you for your credit cards - that was until it was realised that this might mean that the mugger simply takes your finger too! The "arms race" then led to dead finger recognition and the hope that the muggers knew about it!!
And of course, once you yourself have been "chipped", you (or whoever was carrying your arm) could just breeze in and out of the store without any interaction at all other than picking up your supplies. grizelda, Sat, 5th May 2012