WEBVTT 00:00:01.120 --> 00:00:03.910 PROFESSOR: Good morning. 00:00:03.910 --> 00:00:05.570 I'm often hungry for an audience, 00:00:05.570 --> 00:00:07.375 so I came here with eight presentations. 00:00:07.375 --> 00:00:09.460 So I've condensed it down to three, 00:00:09.460 --> 00:00:11.380 which have been merged here, and I'll 00:00:11.380 --> 00:00:15.160 try to be through my section in 10 minutes 00:00:15.160 --> 00:00:18.550 because I have some very other-- 00:00:18.550 --> 00:00:20.990 very knowledgeable people to speak after me. 00:00:20.990 --> 00:00:24.205 I was supposed to speak about evolution of RFID systems. 00:00:30.720 --> 00:00:34.860 And I was trying to capture the notion 00:00:34.860 --> 00:00:37.980 of how these systems evolved under the various constraints 00:00:37.980 --> 00:00:40.620 that have regulated the revolution 00:00:40.620 --> 00:00:43.590 and how they might evolve under future research. 00:00:43.590 --> 00:00:46.470 And there won't be any time to say much about what 00:00:46.470 --> 00:00:48.450 we're doing at Adelaide. 00:00:48.450 --> 00:00:52.020 The sorts of topics that I thought we might collectively 00:00:52.020 --> 00:00:55.710 talk about are here, something on RFID regulations, 00:00:55.710 --> 00:00:58.140 something about antennas, something 00:00:58.140 --> 00:01:02.500 about propagation and protocols, and higher functionality tags. 00:01:02.500 --> 00:01:04.470 Now, I have other speakers that are very 00:01:04.470 --> 00:01:05.920 good on some of those issues. 00:01:05.920 --> 00:01:08.880 So I'm going to talk a little bit about antennas 00:01:08.880 --> 00:01:11.190 and also propagation studies. 00:01:11.190 --> 00:01:14.417 And I'll move to that almost immediately. 00:01:17.220 --> 00:01:20.150 So talking about antenna issues, I'll 00:01:20.150 --> 00:01:22.490 say a little bit about electromagnetic theory 00:01:22.490 --> 00:01:25.670 and maybe something about how antennas work, largely 00:01:25.670 --> 00:01:28.880 through diagrams, talk about near and far fields, 00:01:28.880 --> 00:01:32.660 and talk about what I think are important conclusions that you 00:01:32.660 --> 00:01:36.230 might draw about what you can do with antenna in the near field 00:01:36.230 --> 00:01:37.365 and the far field. 00:01:37.365 --> 00:01:38.990 I think there'll be time for me to talk 00:01:38.990 --> 00:01:42.260 a little bit about the [INAUDIBLE] limit on efficiency 00:01:42.260 --> 00:01:46.160 and maybe show you a couple of simple tag designs. 00:01:46.160 --> 00:01:48.200 I never give a presentation without showing 00:01:48.200 --> 00:01:51.500 this slide, which shows an encapsulation of Maxwell's 00:01:51.500 --> 00:01:53.480 equations in the source and vortex 00:01:53.480 --> 00:01:56.570 interpretation of Helmholtz, which 00:01:56.570 --> 00:01:59.540 I think the more you look at it, the more you realize that it 00:01:59.540 --> 00:02:02.330 contains the secret of life. 00:02:02.330 --> 00:02:04.430 But it boils down to these pictures, 00:02:04.430 --> 00:02:08.780 which show you the source nature of an electric field coming 00:02:08.780 --> 00:02:10.430 from a charge distribution. 00:02:10.430 --> 00:02:12.530 Also shows you the boundary conditions 00:02:12.530 --> 00:02:16.370 you have to contend with when it gets near a conducting surface. 00:02:16.370 --> 00:02:19.520 This shows you the vortex nature of a magnetic field caused 00:02:19.520 --> 00:02:22.130 by a current or displacement current. 00:02:22.130 --> 00:02:23.960 And again, the boundary conditions 00:02:23.960 --> 00:02:26.210 you have to contend with when that becomes 00:02:26.210 --> 00:02:28.100 near a magnetic surface. 00:02:28.100 --> 00:02:30.500 I think this is a picture which will show you 00:02:30.500 --> 00:02:33.530 how an electric field might excite an antenna 00:02:33.530 --> 00:02:36.470 and how a magnetic field might excite an antenna. 00:02:36.470 --> 00:02:40.670 Unfortunately, these are really useful for small antennas. 00:02:40.670 --> 00:02:42.260 When the antennas get a bit big, it 00:02:42.260 --> 00:02:46.030 gets to be a bit more complicated than that. 00:02:46.030 --> 00:02:50.680 Well, these are the fields of a small magnetic dipole. 00:02:50.680 --> 00:02:53.440 You should only look at the red and the blue parts, which 00:02:53.440 --> 00:02:57.010 show you that there's one field that diminishes rather slowly 00:02:57.010 --> 00:02:59.110 and is the first power of distance, 00:02:59.110 --> 00:03:02.440 and another field that diminishes rather quickly. 00:03:02.440 --> 00:03:04.150 That's the third power of distance. 00:03:04.150 --> 00:03:07.030 That shows you that there are near fields, which 00:03:07.030 --> 00:03:09.250 is the blue part, or the far field, which 00:03:09.250 --> 00:03:12.700 is the red part that we should think 00:03:12.700 --> 00:03:15.520 about when we're trying to design systems 00:03:15.520 --> 00:03:18.170 to couple to them. 00:03:18.170 --> 00:03:20.470 We're not going to talk about that slide. 00:03:20.470 --> 00:03:24.820 This is a glimpse of how radar engineers work out power 00:03:24.820 --> 00:03:27.700 transfer between antennas. 00:03:27.700 --> 00:03:29.990 And you can see towards the bottom, 00:03:29.990 --> 00:03:33.310 there's a dependence upon wavelength and a dependence 00:03:33.310 --> 00:03:35.870 upon inverse square power of distance. 00:03:35.870 --> 00:03:38.080 And if you look about that and you think about it, 00:03:38.080 --> 00:03:40.330 you say, well, I'm going to do best 00:03:40.330 --> 00:03:43.450 if I have a very long wavelength and that means 00:03:43.450 --> 00:03:45.190 I should be at a low frequency. 00:03:45.190 --> 00:03:48.800 So the question we should ask is, why is that not true. 00:03:48.800 --> 00:03:51.760 And I think you can come to an answer on that question 00:03:51.760 --> 00:03:54.520 by focusing attention on what happens 00:03:54.520 --> 00:03:57.740 when you're close to an antenna, you've got stored energy. 00:03:57.740 --> 00:04:01.510 So I have what I call a near field coupling theory with some 00:04:01.510 --> 00:04:04.240 of those concepts within it, but I 00:04:04.240 --> 00:04:07.810 think these are the significant conclusions. 00:04:07.810 --> 00:04:11.920 An antenna can be characterized by a coupling volume, 00:04:11.920 --> 00:04:14.350 not really an effective area, and it's 00:04:14.350 --> 00:04:18.010 proportional to the third power of its largest 00:04:18.010 --> 00:04:19.329 physical dimension. 00:04:19.329 --> 00:04:24.800 And you can also characterize it by a quality 00:04:24.800 --> 00:04:28.100 factor which tells you how narrowband the antenna is. 00:04:28.100 --> 00:04:30.350 And that, unfortunately, the quality factors 00:04:30.350 --> 00:04:33.080 are inversely proportional to the third power 00:04:33.080 --> 00:04:34.410 of the distance. 00:04:34.410 --> 00:04:37.640 So I think that this gives us a clue as to why we shouldn't 00:04:37.640 --> 00:04:41.360 always go to low frequencies when we're designing antennas 00:04:41.360 --> 00:04:46.568 because the low frequencies have very large betas in the-- 00:04:50.190 --> 00:04:51.830 have very small betas, actually. 00:04:51.830 --> 00:04:55.250 So the quality factors go up and the bandwidth over which 00:04:55.250 --> 00:04:58.280 are untenable work is impossibly small. 00:04:58.280 --> 00:05:00.200 This leads you to conclusions you 00:05:00.200 --> 00:05:02.930 can draw about optimum operating frequencies. 00:05:02.930 --> 00:05:05.900 And it's really the lowest frequency in which your antenna 00:05:05.900 --> 00:05:08.660 will still be efficient, and that often 00:05:08.660 --> 00:05:12.090 happens to be about the UHF region. 00:05:12.090 --> 00:05:13.800 So that's no surprise, of course, 00:05:13.800 --> 00:05:17.370 why there's a lot of tags working at the UHF region 00:05:17.370 --> 00:05:19.590 where range is required. 00:05:19.590 --> 00:05:24.090 We're also recently interested in what the [INAUDIBLE] theorem 00:05:24.090 --> 00:05:27.720 tells us about over what sort of bandwidth you ought to be able 00:05:27.720 --> 00:05:31.860 to make a UHF tag work, and there's the theorem there 00:05:31.860 --> 00:05:37.980 in a few slides here about what it means in terms of making 00:05:37.980 --> 00:05:40.320 yourself a bad match over frequencies you're not 00:05:40.320 --> 00:05:43.380 interested in and a very good match over frequencies you-- 00:05:43.380 --> 00:05:45.360 well, as good a match as you can manage over 00:05:45.360 --> 00:05:47.350 frequencies you're interested in. 00:05:47.350 --> 00:05:50.520 And I think you conclude that if you 00:05:50.520 --> 00:05:52.950 look at the different problems we face, 00:05:52.950 --> 00:05:57.390 like the USA, which has got one bandwidth, the Japanese, which 00:05:57.390 --> 00:06:00.450 has a smaller bandwidth, but a different part of the spectrum, 00:06:00.450 --> 00:06:05.710 or the European countries, we can ask ourselves, 00:06:05.710 --> 00:06:09.400 can we do a good match over those frequency ranges. 00:06:09.400 --> 00:06:12.510 And I think the conclusions are, it 00:06:12.510 --> 00:06:15.660 does depend a bit upon the characteristics of a circuit. 00:06:15.660 --> 00:06:18.720 And you're not troubled by the theorem, I think, 00:06:18.720 --> 00:06:23.070 if your microcircuit has a relatively low impedance 00:06:23.070 --> 00:06:26.190 of 1,000 ohms in parallel with a picofarad. 00:06:26.190 --> 00:06:29.040 But once it starts to become a very low power circuit, 00:06:29.040 --> 00:06:32.010 it's still about one picofarad of input capacitance, 00:06:32.010 --> 00:06:34.560 but less power consumption, it isn't 00:06:34.560 --> 00:06:39.030 practical to make the antenna work optimally 00:06:39.030 --> 00:06:41.310 without some inefficiency. 00:06:41.310 --> 00:06:44.340 So-- but based on those principles, 00:06:44.340 --> 00:06:46.320 we have designed some small antennas 00:06:46.320 --> 00:06:48.360 with simple matching circuits. 00:06:48.360 --> 00:06:52.140 So down at the bottom here, you can see a tag chip, 00:06:52.140 --> 00:06:54.570 and behind it, there's a capacitance 00:06:54.570 --> 00:06:55.680 in parallel with it. 00:06:55.680 --> 00:06:59.160 Up here, there's another capacitance also in parallel 00:06:59.160 --> 00:07:02.790 with that gap, and that provides a reasonable match 00:07:02.790 --> 00:07:08.050 between the circuit and the radiation impedance. 00:07:08.050 --> 00:07:11.380 So that's I think all I want to say about tags, 00:07:11.380 --> 00:07:13.465 and I think about antennas. 00:07:13.465 --> 00:07:15.090 And I think Rich Fletcher will give you 00:07:15.090 --> 00:07:20.230 something substantially more widespread in that. 00:07:20.230 --> 00:07:23.040 But I think I've got time to talk a little bit about higher 00:07:23.040 --> 00:07:24.810 functionality tags. 00:07:24.810 --> 00:07:28.350 And I think the interesting questions to me 00:07:28.350 --> 00:07:32.230 seem, can you merge electronic article surveillance and data 00:07:32.230 --> 00:07:32.730 tags. 00:07:32.730 --> 00:07:35.190 And I think I'm pretty convinced the answer is not 00:07:35.190 --> 00:07:39.660 easily for the reasons that, to turn them off, 00:07:39.660 --> 00:07:42.270 you're going to lower the queue inevitably 00:07:42.270 --> 00:07:44.670 and you can't get the high quality factors that you 00:07:44.670 --> 00:07:47.670 need in EAS tags if you're turning 00:07:47.670 --> 00:07:49.590 on and off the resonance. 00:07:49.590 --> 00:07:52.470 But I think we've also become interested in that second 00:07:52.470 --> 00:07:55.080 topic, turning on battery-operated tags, 00:07:55.080 --> 00:07:57.300 and I just want to show you some simple results 00:07:57.300 --> 00:08:00.060 for both low power consumption circuits 00:08:00.060 --> 00:08:01.950 and what I call zero power consumption 00:08:01.950 --> 00:08:04.380 circuits for those operations. 00:08:04.380 --> 00:08:07.320 This is what you might do if you have a turn-on circuit down 00:08:07.320 --> 00:08:10.450 at the bottom right and you have a label antenna, 00:08:10.450 --> 00:08:12.900 which you want to both resonate, because resonance 00:08:12.900 --> 00:08:16.590 has the desirable properties of magnifying voltages. 00:08:16.590 --> 00:08:20.760 And so you might want to resonate your available induced 00:08:20.760 --> 00:08:25.030 voltage to produce the maximum voltage across the depletion 00:08:25.030 --> 00:08:28.290 layer capacitance of a rectifying diode 00:08:28.290 --> 00:08:33.210 and then use the DC voltage to apply a turn-on circuit. 00:08:33.210 --> 00:08:36.000 There's a couple of contexts here. 00:08:36.000 --> 00:08:37.799 You might want to apply-- 00:08:37.799 --> 00:08:40.120 get about a volt out of the system 00:08:40.120 --> 00:08:43.679 so you can turn a transistor from desaturation 00:08:43.679 --> 00:08:47.580 to conduction, or you might want to just get about 10 millivolts 00:08:47.580 --> 00:08:50.580 so you can operate the input of a very, very low power 00:08:50.580 --> 00:08:52.500 consumption amplifier. 00:08:52.500 --> 00:08:54.900 This shows some experimental work 00:08:54.900 --> 00:08:59.910 that we were doing to reveal the fact that, if you're working 00:08:59.910 --> 00:09:02.700 at relatively low powers, you can get a nice resonance 00:09:02.700 --> 00:09:04.890 curve as you see on the left. 00:09:04.890 --> 00:09:07.470 But as soon as you start to increase the power levels, 00:09:07.470 --> 00:09:10.080 the nonlinear capacitance variation 00:09:10.080 --> 00:09:13.500 with developed voltage of the diode becomes interplay 00:09:13.500 --> 00:09:18.840 and you end up with that kind of resonance curve 00:09:18.840 --> 00:09:21.060 when the right hand side is quite vertical. 00:09:21.060 --> 00:09:24.270 As it goes off resonance, it suddenly drives itself away 00:09:24.270 --> 00:09:27.130 in operating frequency. 00:09:27.130 --> 00:09:30.640 That's the idea of a low power consumption circuit, which 00:09:30.640 --> 00:09:35.170 will consume about 10 nano amps and turn on 00:09:35.170 --> 00:09:36.970 at about a few millivolts. 00:09:36.970 --> 00:09:39.190 And this is a totally different concept 00:09:39.190 --> 00:09:43.030 in which a vibrating magnetic field might 00:09:43.030 --> 00:09:45.820 shake a magnet, which will distort 00:09:45.820 --> 00:09:49.510 the piezomaterial, which will generate about a volt. 00:09:49.510 --> 00:09:52.180 The analysis of that involves things 00:09:52.180 --> 00:09:57.820 like looking at charge, and displacement, and voltage, 00:09:57.820 --> 00:09:59.860 and torque on the device. 00:09:59.860 --> 00:10:02.260 You can relate material properties 00:10:02.260 --> 00:10:04.490 to structural properties if you know 00:10:04.490 --> 00:10:08.920 the dimensions of the structure, and I think eventually 00:10:08.920 --> 00:10:11.800 produce this expression for the turn-on voltage, which 00:10:11.800 --> 00:10:16.270 will allow you to work out that the concept is 00:10:16.270 --> 00:10:20.620 feasible at frequencies of about 100 kilohertz 00:10:20.620 --> 00:10:25.300 and magnetic fields of the kind that you can use to create 00:10:25.300 --> 00:10:27.250 a stored energy in the foyer. 00:10:27.250 --> 00:10:31.100 And the obvious application is theft detection. 00:10:31.100 --> 00:10:34.300 So if I were to try and walk away with this PC 00:10:34.300 --> 00:10:37.450 and it had such a theft detection tag 00:10:37.450 --> 00:10:40.368 based on these principles in it, it would raise an alarm. 00:10:40.368 --> 00:10:41.410 So I thank you very much. 00:10:41.410 --> 00:10:43.840 I now have pleasure in introducing 00:10:43.840 --> 00:10:45.550 my other colleagues. 00:10:45.550 --> 00:10:48.364 [APPLAUSE] 00:10:51.180 --> 00:10:52.924 Yep. 00:10:52.924 --> 00:10:56.030 AUDIENCE: So our next speaker is Dr. Rich Fletcher, 00:10:56.030 --> 00:10:59.330 who's a visiting scientist here with the AutoID Labs 00:10:59.330 --> 00:11:01.910 and was involved back at the Media Labs 00:11:01.910 --> 00:11:06.500 at MIT when Sanjay, and Dave Brock, 00:11:06.500 --> 00:11:10.200 and so on were in the basement over here. 00:11:10.200 --> 00:11:12.230 Rich was working in RFID in the Media 00:11:12.230 --> 00:11:14.870 Labs, which was a little bit more glitzy at the time, 00:11:14.870 --> 00:11:15.787 I believe. 00:11:15.787 --> 00:11:16.620 RICH FLETCHER: Yeah. 00:11:16.620 --> 00:11:18.633 We had more money back then. 00:11:18.633 --> 00:11:19.300 That's for sure. 00:11:22.880 --> 00:11:24.290 All right. 00:11:24.290 --> 00:11:24.790 Oh. 00:11:32.220 --> 00:11:33.200 Hello? 00:11:33.200 --> 00:11:33.700 OK. 00:11:46.960 --> 00:11:49.630 All right. 00:11:49.630 --> 00:11:52.007 I'm happy to be here today and give you-- 00:11:52.007 --> 00:11:53.590 have a chance to tell you a little bit 00:11:53.590 --> 00:11:59.470 about some part of some aspects of RFID and some of my work. 00:11:59.470 --> 00:12:01.750 I am a visiting scientist at MIT, 00:12:01.750 --> 00:12:04.190 but I work also with MIT Media Lab. 00:12:04.190 --> 00:12:06.190 So some of the slides that I'm going to show you 00:12:06.190 --> 00:12:10.480 and some of the pictures are from different projects at MIT 00:12:10.480 --> 00:12:13.230 as well as some of my company projects as well. 00:12:15.830 --> 00:12:20.410 The basic topic or theme of my talk 00:12:20.410 --> 00:12:23.560 is looking at different electromagnetic issues 00:12:23.560 --> 00:12:27.820 and how they vary depending on different frequencies. 00:12:27.820 --> 00:12:29.360 And there's a lot of slides. 00:12:29.360 --> 00:12:31.340 I'm just going to go through it very quickly. 00:12:31.340 --> 00:12:33.490 But just to give you some sort of flavor 00:12:33.490 --> 00:12:36.100 of the different RFID frequencies that are out there 00:12:36.100 --> 00:12:39.050 and some of the issues involved. 00:12:39.050 --> 00:12:42.520 So as we all know, RFID, is the main goal 00:12:42.520 --> 00:12:44.950 is to send out some sort of signal 00:12:44.950 --> 00:12:48.610 from the reader to the tag and to get some sort of response, 00:12:48.610 --> 00:12:52.720 either reflected power or some modulation from the tag, 00:12:52.720 --> 00:12:57.100 to get information or to use this tag as a sensor. 00:12:57.100 --> 00:12:59.890 And obviously, if there's no power or if there's 00:12:59.890 --> 00:13:01.870 problems with electromagnetics, it 00:13:01.870 --> 00:13:05.050 affects either the turn-on of the tag, or the signaling 00:13:05.050 --> 00:13:08.600 is corrupted, or you have some other types of errors. 00:13:08.600 --> 00:13:10.180 I'm going to start by just giving you 00:13:10.180 --> 00:13:13.720 a very brief, fundamental introduction to some 00:13:13.720 --> 00:13:17.950 of the electromagnetic effects that we look at. 00:13:17.950 --> 00:13:20.290 Obviously, we have different sort of reflections 00:13:20.290 --> 00:13:22.210 depending on what frequency you're looking at. 00:13:22.210 --> 00:13:24.200 The signal also spreads in space. 00:13:24.200 --> 00:13:26.450 It's not like a laser that goes in a straight line, 00:13:26.450 --> 00:13:27.850 so you have spreading loss. 00:13:27.850 --> 00:13:30.220 And every time you go through an interface, 00:13:30.220 --> 00:13:32.620 you get different types of loss just 00:13:32.620 --> 00:13:34.420 through the impedance mismatch. 00:13:34.420 --> 00:13:37.900 There's shielding and detuning of the antennas. 00:13:37.900 --> 00:13:40.870 At the higher frequency, RFID, you 00:13:40.870 --> 00:13:43.520 have multipath reflections from other things in the room, 00:13:43.520 --> 00:13:45.580 including the floor. 00:13:45.580 --> 00:13:47.530 When you go through different slits, 00:13:47.530 --> 00:13:49.240 like layers on the pallet, you also 00:13:49.240 --> 00:13:50.740 have other types of interference. 00:13:50.740 --> 00:13:53.860 The wave interferes with itself. 00:13:53.860 --> 00:13:58.187 You probably experience some of that with your cell phone. 00:13:58.187 --> 00:13:59.770 I'm now going to tell you a little bit 00:13:59.770 --> 00:14:03.500 about the different RFID frequencies that exist. 00:14:03.500 --> 00:14:08.260 Some of these are only used today for EAS or antitheft 00:14:08.260 --> 00:14:10.750 applications, so I don't know how familiar 00:14:10.750 --> 00:14:11.780 you are with those. 00:14:11.780 --> 00:14:13.530 But I thought it would be interesting just 00:14:13.530 --> 00:14:15.730 to give you a flavor of that. 00:14:15.730 --> 00:14:19.310 As we all know, the RFID tags come in many shapes and sizes 00:14:19.310 --> 00:14:21.970 and they've been around since the '70s. 00:14:21.970 --> 00:14:25.750 At the very low end, we have, at 77 Hertz, 00:14:25.750 --> 00:14:28.600 which is an extremely low frequency, that's 00:14:28.600 --> 00:14:32.050 used for library books or certain CDs, antitheft tags, 00:14:32.050 --> 00:14:34.210 little strips that are made by 3M. 00:14:34.210 --> 00:14:36.640 The advantages of this is that it's a magnetic material. 00:14:36.640 --> 00:14:38.500 It's a very thin film, a very low cost. 00:14:38.500 --> 00:14:41.440 It's good for making different types of sensors. 00:14:41.440 --> 00:14:43.730 The disadvantage is that it's shorter range, 00:14:43.730 --> 00:14:45.527 and to generate those magnetic fields, 00:14:45.527 --> 00:14:47.860 you generally need a larger antenna and sometimes higher 00:14:47.860 --> 00:14:48.580 power. 00:14:48.580 --> 00:14:51.830 That's a picture of what some of those tags look like. 00:14:51.830 --> 00:14:56.170 I worked in some of these areas to make sensor tags. 00:14:56.170 --> 00:15:00.910 We made item level temperature sensors using these materials. 00:15:00.910 --> 00:15:03.940 Moving up to slightly higher frequency, 00:15:03.940 --> 00:15:08.290 this is your common 125 kilohertz sort of tag. 00:15:08.290 --> 00:15:12.280 It's sort of the classic RFID tag. 00:15:12.280 --> 00:15:14.717 What's great about it is that it penetrates liquids 00:15:14.717 --> 00:15:16.300 and other materials very well, so it's 00:15:16.300 --> 00:15:19.240 used a lot in industrial applications. 00:15:19.240 --> 00:15:21.580 And has a worldwide frequency. 00:15:21.580 --> 00:15:25.480 It's pretty easy to find, to be able to use it 00:15:25.480 --> 00:15:26.590 anywhere in the world. 00:15:26.590 --> 00:15:30.020 The disadvantage is that it's also somewhat shorter range 00:15:30.020 --> 00:15:32.425 and you need larger coils and antennas. 00:15:34.960 --> 00:15:39.020 There are magnetic versions that work in this frequency as well. 00:15:39.020 --> 00:15:40.700 And let's see. 00:15:40.700 --> 00:15:42.540 I'll just give you-- 00:15:42.540 --> 00:15:43.770 I'll play this. 00:15:43.770 --> 00:15:48.090 This is one of the very early demos that I did at MIT. 00:15:48.090 --> 00:15:51.940 This was at the Media Lab in 1995. 00:15:51.940 --> 00:15:55.350 We were looking at using some of these magnetic materials 00:15:55.350 --> 00:15:59.160 as exploring how they could be used for RFID sensors. 00:15:59.160 --> 00:16:04.410 And there's-- we were looking at using it to measure 00:16:04.410 --> 00:16:08.130 the displacement of a piston, and also detect other objects, 00:16:08.130 --> 00:16:10.790 and also, in this case, detect the squeezing of a toy. 00:16:10.790 --> 00:16:13.290 So as you squeeze it, you can see the little figure animates 00:16:13.290 --> 00:16:14.952 over there. 00:16:14.952 --> 00:16:17.160 So there's a lot you can do with just different types 00:16:17.160 --> 00:16:19.050 of magnetic materials. 00:16:19.050 --> 00:16:23.100 This is the more common form of these sort of tags. 00:16:23.100 --> 00:16:27.720 And these are still by far the largest RFID market, 00:16:27.720 --> 00:16:31.560 even today, used for mostly access control, car mobilizers, 00:16:31.560 --> 00:16:32.580 cattle tagging. 00:16:32.580 --> 00:16:35.190 And the cattle tagging actually started-- 00:16:35.190 --> 00:16:41.010 was invented by an MIT here, Mike Beigel, in the late '70s. 00:16:41.010 --> 00:16:43.910 Because this works pretty well in proximity to metal 00:16:43.910 --> 00:16:45.840 and it's pretty robust, we also used 00:16:45.840 --> 00:16:48.840 it to do some of the early smart shelf work back at the Media 00:16:48.840 --> 00:16:51.840 Lab, this was in the late '90s, where 00:16:51.840 --> 00:16:55.140 we had to read tags either through an LCD display 00:16:55.140 --> 00:16:58.050 or through other types of materials. 00:16:58.050 --> 00:17:00.990 Something else which I'll just mention briefly. 00:17:00.990 --> 00:17:04.260 For a brief time while I was at Motorola and also 00:17:04.260 --> 00:17:06.020 in collaboration with the Media Lab, 00:17:06.020 --> 00:17:09.359 we developed capacitively coupled tags. 00:17:09.359 --> 00:17:11.609 Now, most tags work with magnetic fields 00:17:11.609 --> 00:17:14.040 and inductive coupling with coils. 00:17:14.040 --> 00:17:18.060 For a brief time, we did develop capacitively coupled tags. 00:17:18.060 --> 00:17:20.530 And what's nice about this is that it 00:17:20.530 --> 00:17:22.530 uses electric fields instead of magnetic fields, 00:17:22.530 --> 00:17:25.530 but what's nice about it is that you can make a printed antenna 00:17:25.530 --> 00:17:28.050 and it's very flexible and robust. 00:17:28.050 --> 00:17:29.700 It doesn't require soldering. 00:17:29.700 --> 00:17:32.220 You could rip up the antenna and it still works. 00:17:32.220 --> 00:17:34.192 But unfortunately, Motorola, they 00:17:34.192 --> 00:17:36.150 were losing a lot of money in the late '90s due 00:17:36.150 --> 00:17:37.770 to iridium and other projects, so they 00:17:37.770 --> 00:17:39.817 sold their RFID division. 00:17:39.817 --> 00:17:41.400 But this technology is still out there 00:17:41.400 --> 00:17:44.370 and it's a pretty interesting technology. 00:17:46.950 --> 00:17:49.310 Moving up to slightly higher frequencies. 00:17:49.310 --> 00:17:54.990 HF, for example, 13.56 megahertz. 00:17:54.990 --> 00:17:58.720 This frequency-- you need fewer antennas on your coil, 00:17:58.720 --> 00:18:00.600 so you can make a lower cost tag. 00:18:00.600 --> 00:18:02.640 You can make it out of foil, however, 00:18:02.640 --> 00:18:05.070 because the penetration depth is still rather thick. 00:18:05.070 --> 00:18:08.430 You need a thick metal layer. 00:18:08.430 --> 00:18:10.950 And also, it requires a crossover to do a coil. 00:18:10.950 --> 00:18:15.030 So it's not as low cost as, say, a UHF tag. 00:18:15.030 --> 00:18:17.040 But it does enable a very low cost reader. 00:18:17.040 --> 00:18:19.800 You can make extremely low cost readers at this frequency. 00:18:19.800 --> 00:18:22.500 And this is used a lot in toy applications. 00:18:22.500 --> 00:18:25.215 You can buy a reader for a couple of dollars 00:18:25.215 --> 00:18:28.320 in Hong Kong that's used in the toy market. 00:18:28.320 --> 00:18:29.907 And it's also very nice for making 00:18:29.907 --> 00:18:31.740 different types of sensors because the value 00:18:31.740 --> 00:18:36.300 of inductances and capacitances at these frequency range 00:18:36.300 --> 00:18:40.122 is just right for different types of-- 00:18:40.122 --> 00:18:42.330 for integrating it into a tag and for different types 00:18:42.330 --> 00:18:42.830 of sensors. 00:18:45.300 --> 00:18:48.660 This is some of-- another type of reader 00:18:48.660 --> 00:18:54.120 that was developed in my lab for exploring 00:18:54.120 --> 00:18:57.460 how cell phones can be used as a reader and also as a tag. 00:18:57.460 --> 00:19:00.370 So you can transfer data from one cell phone to another. 00:19:00.370 --> 00:19:03.360 You can also load up a variety of IDs onto your phone 00:19:03.360 --> 00:19:04.530 and read it. 00:19:04.530 --> 00:19:07.620 So this sort of near field applications 00:19:07.620 --> 00:19:09.990 is what's possible and convenient to do 00:19:09.990 --> 00:19:11.040 at low frequencies. 00:19:11.040 --> 00:19:13.560 And obviously, another market that's growing very fast 00:19:13.560 --> 00:19:14.790 is payment. 00:19:14.790 --> 00:19:17.300 It's growing all over the world. 00:19:17.300 --> 00:19:20.280 I don't need to really talk about that. 00:19:20.280 --> 00:19:22.840 Moving up to higher frequencies. 00:19:22.840 --> 00:19:25.620 So UHF tags, which is obviously what's 00:19:25.620 --> 00:19:28.030 been getting the most attention for supply chain, 00:19:28.030 --> 00:19:34.140 and this is what's being used in EPC, in the EPC world, mostly. 00:19:34.140 --> 00:19:38.700 It has-- its advantages are a very low cost tag. 00:19:38.700 --> 00:19:41.250 You can now use very thin metal conductors. 00:19:41.250 --> 00:19:43.290 You can make a single layer antenna, 00:19:43.290 --> 00:19:47.110 so it brings the cost of tag way down. 00:19:47.110 --> 00:19:49.410 It also has an increased read range. 00:19:49.410 --> 00:19:52.560 The antennas at this frequency have a larger capture-- 00:19:52.560 --> 00:19:57.720 a larger cross section, which allows you to get a longer 00:19:57.720 --> 00:20:00.430 range for your reader. 00:20:00.430 --> 00:20:02.800 But the disadvantages are that there's null spots. 00:20:02.800 --> 00:20:06.780 Because the wavelengths are on the order of a meter, 00:20:06.780 --> 00:20:11.820 you get null spots on that sort of scale. 00:20:11.820 --> 00:20:13.860 And there are also chipless versions 00:20:13.860 --> 00:20:16.830 of this that work for EAS and I believe 00:20:16.830 --> 00:20:22.410 Professor Cole was involved in an early version of this. 00:20:22.410 --> 00:20:25.620 The cost of UHF technology in general 00:20:25.620 --> 00:20:31.030 has been plummeting, which is pretty amazing. 00:20:31.030 --> 00:20:36.960 The tags-- you can now buy tags for less than $0.20 in even low 00:20:36.960 --> 00:20:40.770 quantities, such as 1,000 tags. 00:20:40.770 --> 00:20:43.640 The cost of readers, due to the advent 00:20:43.640 --> 00:20:46.850 of wireless technologies in commercial products 00:20:46.850 --> 00:20:50.600 and consumer electronics, the cost of CMOS radio ICs 00:20:50.600 --> 00:20:53.130 has been coming down dramatically. 00:20:53.130 --> 00:20:54.770 And you can now build a reader-- 00:20:54.770 --> 00:20:56.395 there's a reader I designed last summer 00:20:56.395 --> 00:21:01.040 with just a parts cost of $30 and it could read EPC gen 2 00:21:01.040 --> 00:21:07.340 and gen 1 and with a read range of a couple of meters. 00:21:07.340 --> 00:21:10.880 So moving up to microwave. 00:21:10.880 --> 00:21:17.570 So microwave, and particular 2.4 gigahertz or the ISM band, this 00:21:17.570 --> 00:21:21.860 is very attractive because it's a smaller antenna. 00:21:21.860 --> 00:21:24.290 It's a worldwide frequency. 00:21:24.290 --> 00:21:26.540 There's many other standards and wireless technologies 00:21:26.540 --> 00:21:27.270 that work there. 00:21:27.270 --> 00:21:29.430 So you have the economy of scale. 00:21:29.430 --> 00:21:31.820 So the parts and the antennas are already available 00:21:31.820 --> 00:21:35.690 and you can make RFID systems in this frequency 00:21:35.690 --> 00:21:38.100 very cheaply as well. 00:21:38.100 --> 00:21:41.630 But disadvantages, as with UHF, is that it's easily shielded. 00:21:41.630 --> 00:21:45.620 Here's an example of a ZigBee tag that I designed. 00:21:45.620 --> 00:21:50.330 And one big advantage here is just that, obviously, this 00:21:50.330 --> 00:21:51.320 is battery powered. 00:21:51.320 --> 00:21:55.250 However, you can make a very tiny reader 00:21:55.250 --> 00:21:56.660 that's very low cost. 00:21:56.660 --> 00:21:59.150 And the advantage, for example, for a pallet tracking 00:21:59.150 --> 00:22:01.565 application is rather than having 00:22:01.565 --> 00:22:04.280 a $1,000 reader at each portal, you 00:22:04.280 --> 00:22:10.410 can now have one reader covering 10 dock doors 00:22:10.410 --> 00:22:13.220 at a fraction of the cost. 00:22:13.220 --> 00:22:17.300 Then moving up finally to the next frequency range, 00:22:17.300 --> 00:22:21.470 you have a millimeter wave or higher microwaves. 00:22:21.470 --> 00:22:25.350 And these are very tiny dipoles, very tiny antennas. 00:22:25.350 --> 00:22:28.940 And this is mostly used today for anticounterfeiting 00:22:28.940 --> 00:22:31.460 because they can embed the materials into things 00:22:31.460 --> 00:22:35.510 like passports, or fabrics, or other printed media. 00:22:35.510 --> 00:22:37.640 You could-- because the antennas are so small, 00:22:37.640 --> 00:22:39.530 you can make an array of the antennas 00:22:39.530 --> 00:22:42.350 and you can use-- you can make a phased array so you 00:22:42.350 --> 00:22:44.030 can steer your beam around. 00:22:44.030 --> 00:22:48.080 And here's a picture of a little reader, 00:22:48.080 --> 00:22:51.890 and you can see how tiny that antenna array is. 00:22:51.890 --> 00:22:55.123 There's a 26-- this particular one is at 26 and 1/2 gigahertz. 00:22:55.123 --> 00:22:56.540 And what you see in the background 00:22:56.540 --> 00:22:57.980 are some of the printed dipoles. 00:22:57.980 --> 00:23:01.100 This happens to be printed on a polyester sheet, 00:23:01.100 --> 00:23:03.380 but it comes in many other forms and it's 00:23:03.380 --> 00:23:06.720 used for food packaging and other things around the world. 00:23:06.720 --> 00:23:11.900 So in anticounterfeiting in currency and so forth. 00:23:11.900 --> 00:23:15.093 Just to tell you briefly of Leena, the next speaker, 00:23:15.093 --> 00:23:17.510 is going to give you an example of some of the work that's 00:23:17.510 --> 00:23:19.850 being done here in electromagnetics at the AutoID 00:23:19.850 --> 00:23:22.430 Lab, but I just wanted to mention the topics 00:23:22.430 --> 00:23:24.980 that we generally look at that fall 00:23:24.980 --> 00:23:27.290 under the category of electromagnetics 00:23:27.290 --> 00:23:31.880 is reader antenna design and also tag antenna design. 00:23:31.880 --> 00:23:35.540 But things like geometry, the materials interaction, 00:23:35.540 --> 00:23:39.690 and the overall propagation between the reader and the tag. 00:23:39.690 --> 00:23:42.787 So one important point I'd like to make here 00:23:42.787 --> 00:23:44.870 is that it's important to standardize all of that, 00:23:44.870 --> 00:23:49.070 not just the protocol and the reader. 00:23:49.070 --> 00:23:51.590 I'm going to fast forward through some of this. 00:23:51.590 --> 00:23:53.007 Some of the things that we've done 00:23:53.007 --> 00:23:56.120 is we've built simulation tools. 00:23:56.120 --> 00:23:58.070 Some of you have seen this already. 00:23:58.070 --> 00:24:02.690 We've built probes that we could use 00:24:02.690 --> 00:24:06.500 to-- it's a semi active or semi passive tag 00:24:06.500 --> 00:24:10.460 that we can embed inside a pallet that takes readings 00:24:10.460 --> 00:24:12.740 of the field, it samples a field, 00:24:12.740 --> 00:24:16.980 and it talks back to the reader using the EPC protocol. 00:24:16.980 --> 00:24:20.750 So you can sprinkle some of these in with standard tags 00:24:20.750 --> 00:24:23.810 to give you more information about your reader installation. 00:24:23.810 --> 00:24:27.470 And we've done a variety of propagation studies looking 00:24:27.470 --> 00:24:31.280 at how the different thicknesses, materials, 00:24:31.280 --> 00:24:32.960 and different properties of the material 00:24:32.960 --> 00:24:35.480 affect the read rates and the propagation. 00:24:35.480 --> 00:24:40.920 And, well, surprise, surprise, Maxwell's equations works. 00:24:40.920 --> 00:24:45.230 And so we've also looked at certain geometries 00:24:45.230 --> 00:24:47.360 for pallet stacking. 00:24:47.360 --> 00:24:48.930 How we can use printed inks and try 00:24:48.930 --> 00:24:50.840 to look at low cost implementations, 00:24:50.840 --> 00:24:52.940 how packaging can be-- 00:24:52.940 --> 00:24:54.650 the proper packaging design can be 00:24:54.650 --> 00:24:57.170 used to improve the read rates. 00:24:57.170 --> 00:25:00.170 I could tell you more detail in person if you're interested. 00:25:00.170 --> 00:25:04.820 And finally, we've looked at how the evolution of packaging 00:25:04.820 --> 00:25:07.670 over time, how, starting with slap and ship, 00:25:07.670 --> 00:25:13.250 you can be smarter about the way you place the tags and the way 00:25:13.250 --> 00:25:15.440 you fabricate your boxes. 00:25:15.440 --> 00:25:18.110 Maybe eventually, if we get to the point 00:25:18.110 --> 00:25:21.800 where the tags are actually embedded into the cardboard 00:25:21.800 --> 00:25:25.610 boxes, we can vastly improve the read rates. 00:25:25.610 --> 00:25:28.460 And your ROI level depends on your application 00:25:28.460 --> 00:25:30.900 and depends on the particular company. 00:25:30.900 --> 00:25:35.240 So in conclusion, as Sanjay and everybody else has said, 00:25:35.240 --> 00:25:37.830 there's a lot of work to do. 00:25:37.830 --> 00:25:40.190 But I just wanted to mention that there's 00:25:40.190 --> 00:25:47.070 much more to do than just protocol and tag IC design. 00:25:47.070 --> 00:25:51.750 And there's also-- the EPC sort of RFID 00:25:51.750 --> 00:25:55.770 is just a very small slice of the potential RFID technologies 00:25:55.770 --> 00:25:58.690 and frequencies that are available out there. 00:25:58.690 --> 00:26:03.180 So there's a lot that we can look at as well. 00:26:03.180 --> 00:26:06.640 I think I'll end there and we'll move on. 00:26:06.640 --> 00:26:07.140 Thank you. 00:26:07.140 --> 00:26:09.590 [APPLAUSE] 00:26:13.127 --> 00:26:14.460 STEPHEN GRAVES: Thank you, Rich. 00:26:14.460 --> 00:26:17.640 So our next speaker-- we thought since this conference 00:26:17.640 --> 00:26:20.250 was about academic collaboration that we 00:26:20.250 --> 00:26:23.550 might share with you the results of a collaboration 00:26:23.550 --> 00:26:27.600 between Tampere University of Technology 00:26:27.600 --> 00:26:32.580 and the Rama Research Unit and the Auto-ID Labs. 00:26:32.580 --> 00:26:37.590 And Leena Ukkonen is here to share with us 00:26:37.590 --> 00:26:42.060 some specific work in the antenna-design area 00:26:42.060 --> 00:26:44.640 that she's worked on over the course of the last year 00:26:44.640 --> 00:26:46.352 between the two institutions. 00:27:17.160 --> 00:27:19.920 LEENA UKKONEN: So good afternoon, everyone, and it's 00:27:19.920 --> 00:27:26.060 great to be here today talking about the research on antenna 00:27:26.060 --> 00:27:27.240 designs. 00:27:27.240 --> 00:27:33.580 And we have been collaborating with the MIT IDEA Lab 00:27:33.580 --> 00:27:35.880 since 2001. 00:27:35.880 --> 00:27:42.240 And our collaboration began through the Auto-ID Center 00:27:42.240 --> 00:27:44.960 Ergonomic Alliance. 00:27:44.960 --> 00:27:50.150 And, well, the main research focus 00:27:50.150 --> 00:27:53.300 on RFID at our resource institute 00:27:53.300 --> 00:27:58.010 in Tampere University of Technology Rama Research Unit 00:27:58.010 --> 00:27:59.630 is tag antennas. 00:27:59.630 --> 00:28:05.390 And I was here two years ago as a visiting PhD student working 00:28:05.390 --> 00:28:11.600 on tag antennas for challenging objects like objects containing 00:28:11.600 --> 00:28:12.860 metals and liquids. 00:28:12.860 --> 00:28:18.260 And this collaboration has been continuing also 00:28:18.260 --> 00:28:20.630 after when I went back to Finland, 00:28:20.630 --> 00:28:23.990 and it's been very fruitful and great to work 00:28:23.990 --> 00:28:27.050 with the MIT Auto-ID Labs. 00:28:27.050 --> 00:28:32.190 And about today's presentation-- 00:28:32.190 --> 00:28:37.920 well, since the main research focus of our lab 00:28:37.920 --> 00:28:43.660 is tag antennas, I'm talking about omnidirectional tag 00:28:43.660 --> 00:28:47.310 antenna for passive UHF RFID of paper reels. 00:28:47.310 --> 00:28:51.120 And this has been our main research topic 00:28:51.120 --> 00:28:57.250 at Rama Research Unit in 2005. 00:28:57.250 --> 00:28:59.710 And first I'm going to tell something 00:28:59.710 --> 00:29:04.450 about the challenges in applying passive UHF RFID in paper 00:29:04.450 --> 00:29:06.850 industry because at the moment there 00:29:06.850 --> 00:29:10.930 is an urgent need in paper industry for an identification 00:29:10.930 --> 00:29:14.980 system that would carry on the identification code 00:29:14.980 --> 00:29:17.620 throughout the whole supply chain of the reel 00:29:17.620 --> 00:29:21.400 because at the moment when all the bar-code systems are used, 00:29:21.400 --> 00:29:25.480 they are placed on the surface of the reel, on the wrapping, 00:29:25.480 --> 00:29:28.870 and all the identification is removed 00:29:28.870 --> 00:29:30.400 when the wrapping is removed. 00:29:30.400 --> 00:29:34.000 And then you cannot know anymore like the origin of the reel 00:29:34.000 --> 00:29:37.720 and that kind of things, which would be important-- 00:29:37.720 --> 00:29:40.820 for example, printing companies. 00:29:40.820 --> 00:29:43.750 So that's why in our approach we are 00:29:43.750 --> 00:29:48.400 placing a tag on the paper reel core under the wrapped paper. 00:29:48.400 --> 00:29:54.800 And that has a lot of effects on the RFID system performance 00:29:54.800 --> 00:29:56.800 and the tag-antenna performance that 00:29:56.800 --> 00:30:00.130 have to be taken into account. 00:30:00.130 --> 00:30:04.540 And, well, of course because we are operating 00:30:04.540 --> 00:30:07.540 in the industrial environment, there 00:30:07.540 --> 00:30:09.430 is all kind of background noise that 00:30:09.430 --> 00:30:13.450 has to be taken into account, and also the environment 00:30:13.450 --> 00:30:17.090 can be kind of hard-- for example, cold environments 00:30:17.090 --> 00:30:18.350 and such. 00:30:18.350 --> 00:30:22.270 And of course there are a lot of different paper qualities 00:30:22.270 --> 00:30:24.820 and cardboard, and we would like to have 00:30:24.820 --> 00:30:29.350 a tag that would function with all of those different paper 00:30:29.350 --> 00:30:33.140 qualities and also with cardboard. 00:30:33.140 --> 00:30:36.200 But the biggest challenge so far has 00:30:36.200 --> 00:30:40.620 been developing an omnidirectional tag antenna 00:30:40.620 --> 00:30:42.720 which is indisplaceable-- 00:30:42.720 --> 00:30:45.140 for example, in lift-truck handling. 00:30:45.140 --> 00:30:50.240 Because as you can see from here, the guy who's 00:30:50.240 --> 00:30:53.690 driving this truck, he just grabs 00:30:53.690 --> 00:30:56.720 the reel, and the identification, 00:30:56.720 --> 00:31:00.470 which is carried out using a radar unit that 00:31:00.470 --> 00:31:03.140 would be integrated into this truck, 00:31:03.140 --> 00:31:06.920 it has to be automatic so the driver doesn't 00:31:06.920 --> 00:31:10.070 have to look for any direction where the tag is. 00:31:10.070 --> 00:31:15.050 It has to be able to be read omnidirectional 00:31:15.050 --> 00:31:16.580 around the reel. 00:31:16.580 --> 00:31:20.040 And, of course, in general in paper-mill environment, 00:31:20.040 --> 00:31:24.170 if you just can identify the antenna with only one reader 00:31:24.170 --> 00:31:26.720 antenna, that would be also good anyway. 00:31:29.620 --> 00:31:33.190 Well, next there is something about this concept 00:31:33.190 --> 00:31:35.830 of omnidirectional reading and also 00:31:35.830 --> 00:31:38.180 of the structure of this paper reel. 00:31:38.180 --> 00:31:42.220 So here you can see the vertically orientated reel. 00:31:42.220 --> 00:31:45.190 And first there is a reel core which is 00:31:45.190 --> 00:31:48.400 fabricated of hard cardboard. 00:31:48.400 --> 00:31:51.640 And then the tag is placed on the core, 00:31:51.640 --> 00:31:54.680 and then the paper is wrapped around the core. 00:31:54.680 --> 00:31:58.780 And typically these thicknesses of the paper layer 00:31:58.780 --> 00:32:02.950 that's wrapped around the core varies between 500 and 600 00:32:02.950 --> 00:32:04.630 millimeters. 00:32:04.630 --> 00:32:08.230 And also the length of the reel can vary, 00:32:08.230 --> 00:32:12.400 or it varies from something like 300 millimeters up 00:32:12.400 --> 00:32:15.280 to 2.5 meters. 00:32:15.280 --> 00:32:17.650 And, well, the omnidirectional reading 00:32:17.650 --> 00:32:22.240 means that the paper reel or the tag 00:32:22.240 --> 00:32:27.550 can be identified omnidirectionally 360 degrees 00:32:27.550 --> 00:32:28.810 around the reel. 00:32:28.810 --> 00:32:33.820 So you don't have to care where the tag has 00:32:33.820 --> 00:32:36.010 been placed on this core. 00:32:36.010 --> 00:32:40.500 So you can read a tag around the reel. 00:32:40.500 --> 00:32:46.170 And, well, in this tag-creation process, there are many steps, 00:32:46.170 --> 00:32:50.170 and I'm going to briefly describe them. 00:32:50.170 --> 00:32:52.860 So first there is modeling. 00:32:52.860 --> 00:32:56.940 This picture has been taken from the simulation software that is 00:32:56.940 --> 00:32:59.180 based on finite element method. 00:32:59.180 --> 00:33:03.760 And you can see that is a very real-life structure. 00:33:03.760 --> 00:33:05.855 So there's the reel. 00:33:05.855 --> 00:33:08.850 The tag is placed on the core, the reel core. 00:33:08.850 --> 00:33:12.400 And then the paper is wrapped around the core. 00:33:12.400 --> 00:33:16.510 So it's like a real industrial paper reel. 00:33:16.510 --> 00:33:20.350 And here are some radiation patterns 00:33:20.350 --> 00:33:26.140 of those tag-antenna models that we've been 00:33:26.140 --> 00:33:28.920 modeling during this project. 00:33:28.920 --> 00:33:33.280 So those pictures on the left are some earlier stages 00:33:33.280 --> 00:33:34.850 of modeling. 00:33:34.850 --> 00:33:41.950 So because we try different antenna geometries 00:33:41.950 --> 00:33:49.600 and tried changing some things and some geometries 00:33:49.600 --> 00:33:53.500 on those designs affects the radiation pattern. 00:33:53.500 --> 00:33:57.280 And the bigger one on the left is 00:33:57.280 --> 00:34:00.380 the stage we are at the moment. 00:34:00.380 --> 00:34:06.070 So you can see that the tag antenna radiates 00:34:06.070 --> 00:34:11.630 into all directions around the reel. 00:34:11.630 --> 00:34:15.889 And, well, then there is also, of course, measurements. 00:34:15.889 --> 00:34:20.510 And these pictures here are just some basic measurement setups 00:34:20.510 --> 00:34:26.540 with network analyzer to see how adding paper on a tag 00:34:26.540 --> 00:34:30.420 affects, for example, resonance frequency. 00:34:30.420 --> 00:34:33.480 And, well, because our goal was to develop 00:34:33.480 --> 00:34:36.060 an omnidirectional tag antenna, we 00:34:36.060 --> 00:34:38.790 developed an omnidirectional model 00:34:38.790 --> 00:34:43.110 with which we can test the omnidirectional reading 00:34:43.110 --> 00:34:46.260 of the tags because when we go to the paper mill, 00:34:46.260 --> 00:34:50.710 we want to have as good tags as possible. 00:34:50.710 --> 00:34:53.760 So we developed this model, and we 00:34:53.760 --> 00:34:56.429 could test the omnidirectional reading 00:34:56.429 --> 00:34:58.960 from all the directions. 00:34:58.960 --> 00:35:06.210 So we have 16 measurement points and four different distances. 00:35:06.210 --> 00:35:11.310 And if the tag was identified at all the directions at all 00:35:11.310 --> 00:35:14.490 those distances, it was omnidirectional 00:35:14.490 --> 00:35:18.120 also inside a real paper reel. 00:35:18.120 --> 00:35:21.970 And here is the omnidirectional tag antenna, 00:35:21.970 --> 00:35:24.390 which we call the C tag, because when 00:35:24.390 --> 00:35:29.490 it's wrapped around this core, it formed a shape of C. 00:35:29.490 --> 00:35:32.430 And here you can see the antenna design flat. 00:35:32.430 --> 00:35:37.740 And, well, here it is in the paper-mill environment 00:35:37.740 --> 00:35:40.860 mounted on the core before the paper is wrapped around it. 00:35:43.440 --> 00:35:48.000 And we did a lot of practical testing 00:35:48.000 --> 00:35:50.340 with this antenna design, and here you 00:35:50.340 --> 00:35:53.430 can see our measurement equipment. 00:35:53.430 --> 00:35:59.100 And we used Alien Technologies European reader unit, Alien 00:35:59.100 --> 00:36:02.070 Technology straps as a microchip. 00:36:02.070 --> 00:36:06.930 And the reader was based on new ETSI regulations, 00:36:06.930 --> 00:36:14.090 and it had two-watt ERP transmitting power. 00:36:14.090 --> 00:36:18.020 And here are some pictures from the measurements. 00:36:18.020 --> 00:36:21.700 So this is basically how the read ranges were measured 00:36:21.700 --> 00:36:25.750 in a paper mill, and we could move the reader 00:36:25.750 --> 00:36:29.050 units and the reader antenna and roll the reel on the floor 00:36:29.050 --> 00:36:33.430 so that we could measure the omnidirectional reading. 00:36:33.430 --> 00:36:35.680 And here is some more pictures, and here you 00:36:35.680 --> 00:36:39.310 can see how the tag is placed on the core. 00:36:39.310 --> 00:36:43.910 And we didn't actually kill any tags on this process, 00:36:43.910 --> 00:36:46.570 so that was kind of good news that it 00:36:46.570 --> 00:36:49.480 went through the process. 00:36:49.480 --> 00:36:53.690 And, well, now I'm moving on to the measurement results. 00:36:53.690 --> 00:36:57.220 So we measured in the paper mill coated printing paper 00:36:57.220 --> 00:37:02.890 with reel diameters varying from 1,200 to 1,300 millimeters. 00:37:02.890 --> 00:37:07.600 And here is the data of the read ranges 00:37:07.600 --> 00:37:10.250 that were achieved around the reel. 00:37:10.250 --> 00:37:13.510 So here you can see that it's read omnidirectionally. 00:37:13.510 --> 00:37:16.960 And the read ranges have some variation 00:37:16.960 --> 00:37:20.500 around the reel, which was also expected from the simulation 00:37:20.500 --> 00:37:23.210 results. 00:37:23.210 --> 00:37:26.780 And, well, I'm going to briefly tell something 00:37:26.780 --> 00:37:30.120 about identification of cardboard reels. 00:37:30.120 --> 00:37:35.120 And, well, they said that it's been impossible 00:37:35.120 --> 00:37:37.110 with conventional tags. 00:37:37.110 --> 00:37:42.080 So we tested our tag also for cardboard identification. 00:37:42.080 --> 00:37:44.270 And there is some more challenges 00:37:44.270 --> 00:37:46.010 with these cardboard reels, which 00:37:46.010 --> 00:37:50.330 are larger diameter and also the more layered and inhomogeneous 00:37:50.330 --> 00:37:52.790 structure of the cardboard, which increases 00:37:52.790 --> 00:37:55.670 the boundary-length effect. 00:37:55.670 --> 00:37:58.940 So we did some practical testing, 00:37:58.940 --> 00:38:02.120 and the tag antenna was the same that we 00:38:02.120 --> 00:38:05.450 used with the paper reel. 00:38:05.450 --> 00:38:09.800 So it was not yet optimized for the cardboard. 00:38:09.800 --> 00:38:13.280 So the goal of the first testing was 00:38:13.280 --> 00:38:19.730 to identify the tag through the cardboard reel, which 00:38:19.730 --> 00:38:23.840 we did, as you can see here. 00:38:23.840 --> 00:38:30.800 And also, yeah, kind of the most surprising and good result 00:38:30.800 --> 00:38:38.870 was that we could identify the reel 180 degrees around it. 00:38:38.870 --> 00:38:40.910 So we achieved better results. 00:38:40.910 --> 00:38:48.020 That was like the goal of these first testing. 00:38:48.020 --> 00:38:52.470 Well, there is still a lot of work to do on this. 00:38:52.470 --> 00:38:54.830 But to our best knowledge, this is 00:38:54.830 --> 00:38:59.270 the first omnidirectional tag antenna for passive UHF RFID 00:38:59.270 --> 00:39:02.390 paper reels that can be read omnidirectionally 00:39:02.390 --> 00:39:05.360 with standardized RFID equipment. 00:39:05.360 --> 00:39:08.690 And it has been tested with copy paper and coated printing 00:39:08.690 --> 00:39:11.690 paper, and also the cardboard has 00:39:11.690 --> 00:39:17.150 been tested with 180 degree identification around the reel. 00:39:17.150 --> 00:39:21.050 So in the future, we will develop an omnidirectional tag 00:39:21.050 --> 00:39:24.810 antenna also for cardboard reels. 00:39:24.810 --> 00:39:28.080 And we will test and develop the antenna 00:39:28.080 --> 00:39:32.580 also for American and Asian UHF RFID bands 00:39:32.580 --> 00:39:35.490 so that we could achieve a global tag that 00:39:35.490 --> 00:39:38.880 could be used around the world. 00:39:38.880 --> 00:39:43.510 And also longer read ranges will be achieved. 00:39:43.510 --> 00:39:47.130 So we've talked about this with industry people, 00:39:47.130 --> 00:39:52.020 and they say that a minimum of 0.5 meters from the paper 00:39:52.020 --> 00:39:55.590 surface would be required. 00:39:55.590 --> 00:39:58.020 And also the tag has to be evaluated 00:39:58.020 --> 00:40:03.530 in harsh environments-- for example, in cold temperatures. 00:40:03.530 --> 00:40:09.350 And, well, there is some other research project also in 2006. 00:40:09.350 --> 00:40:12.380 So we'll continue also developing the tag antenna 00:40:12.380 --> 00:40:15.140 for metallic- and liquid-containing objects. 00:40:15.140 --> 00:40:20.000 So basically we will continue on the miniaturization 00:40:20.000 --> 00:40:22.500 of the [INAUDIBLE] badge-type tag. 00:40:22.500 --> 00:40:25.620 So thank you for your attention. 00:40:25.620 --> 00:40:26.350 Thank you. 00:40:26.350 --> 00:40:30.086 [APPLAUSE] 00:40:30.355 --> 00:40:31.730 STEPHEN GRAVES: Thank you, Leena. 00:40:31.730 --> 00:40:35.270 And our final speaker on this panel is Dr. Alan Levesque. 00:40:35.270 --> 00:40:38.720 He's a colleague of Dr. Kaveh Pahlavan 00:40:38.720 --> 00:40:42.110 at the Center for Wireless Information Network Studies. 00:40:42.110 --> 00:40:46.728 Kaveh was, I think, a chair and an IEEE Wi-Fi committee at one 00:40:46.728 --> 00:40:47.770 point or something, but-- 00:40:47.770 --> 00:40:48.562 ALAN LEVESQUE: Yes. 00:40:48.562 --> 00:40:53.060 Actually Kaveh's been involved in wireless issues 00:40:53.060 --> 00:40:56.120 since the early days of wireless LANs. 00:40:56.120 --> 00:41:01.160 Those of you that know the business at all in the greater 00:41:01.160 --> 00:41:04.490 Boston area know that about 15 years ago-- 00:41:04.490 --> 00:41:05.570 I like that slide. 00:41:05.570 --> 00:41:10.010 Someone said hope to hype to implementation. 00:41:10.010 --> 00:41:11.030 I like that. 00:41:11.030 --> 00:41:14.840 About 15 years ago was the hype phase of wireless LANs, 00:41:14.840 --> 00:41:20.090 and a lot of that activity was a number of startup companies 00:41:20.090 --> 00:41:21.840 actually in Massachusetts. 00:41:21.840 --> 00:41:24.470 And my colleague, Kaveh Pahlavan, 00:41:24.470 --> 00:41:27.760 was actually involved with several of those startups. 00:41:27.760 --> 00:41:28.510 So that's a good-- 00:41:28.510 --> 00:41:30.177 STEPHEN GRAVES: But in any case, they're 00:41:30.177 --> 00:41:32.830 doing some very interesting work in location-based tracking 00:41:32.830 --> 00:41:35.632 that we thought would complement this session nicely. 00:41:35.632 --> 00:41:36.590 ALAN LEVESQUE: Exactly. 00:41:36.590 --> 00:41:37.910 Thank you, Steve. 00:41:37.910 --> 00:41:41.990 Steve's provided-- I should do that, yes. 00:41:46.970 --> 00:41:48.500 Thank you, Steve. 00:41:48.500 --> 00:41:54.230 Steve has provided a nice introduction for me. 00:42:00.620 --> 00:42:03.230 How do we move to the next presentation? 00:42:23.070 --> 00:42:24.070 Good. 00:42:24.070 --> 00:42:24.842 Thank you. 00:42:24.842 --> 00:42:25.990 Very good. 00:42:25.990 --> 00:42:28.330 I'm painfully aware that I'm the one who's 00:42:28.330 --> 00:42:30.460 keeping us all from lunch now, so I'm 00:42:30.460 --> 00:42:37.960 going to do some real-time editing as I go along here. 00:42:37.960 --> 00:42:40.220 Kaveh, in fact, intended to be here, 00:42:40.220 --> 00:42:42.940 but he has a commitment in Japan. 00:42:42.940 --> 00:42:46.300 So based on the weather report that Steve gave this morning 00:42:46.300 --> 00:42:49.990 about Japan, he may have swapped a snowstorm in Cambridge 00:42:49.990 --> 00:42:51.850 for a snowstorm in Japan. 00:42:51.850 --> 00:42:54.170 I'm not sure. 00:42:54.170 --> 00:43:02.080 As Steve said, our emphasis in the last few years 00:43:02.080 --> 00:43:08.200 in the wireless center at WPI has been on location sensing. 00:43:08.200 --> 00:43:10.810 We also use the term localization. 00:43:10.810 --> 00:43:17.560 And because of some of the previous presentations 00:43:17.560 --> 00:43:20.680 and obviously the background of knowledge that many of you 00:43:20.680 --> 00:43:25.810 will have, I'll be able to skip over some of this. 00:43:25.810 --> 00:43:29.380 We have actually been focused in recent years 00:43:29.380 --> 00:43:34.780 primarily on public safety and military applications, 00:43:34.780 --> 00:43:41.090 partly sponsored by DARPA, by NSF, and with some membership 00:43:41.090 --> 00:43:44.170 subscription-type sponsorship from member 00:43:44.170 --> 00:43:46.585 companies in the center. 00:43:50.650 --> 00:43:53.170 I'll leave this up long enough to point out 00:43:53.170 --> 00:43:56.200 that several people made mention of the history 00:43:56.200 --> 00:44:00.040 of this technology going back 50 years to World War II, 00:44:00.040 --> 00:44:02.380 and everyone has a little bit different take on it. 00:44:02.380 --> 00:44:06.070 My take is the fact that that era introduced 00:44:06.070 --> 00:44:09.730 the use of what are called technically net broadcast 00:44:09.730 --> 00:44:11.830 radios, push-to-talk radios. 00:44:11.830 --> 00:44:14.560 The devices were so-called walkie talkies 00:44:14.560 --> 00:44:16.870 about the weight of a brick and about the volume 00:44:16.870 --> 00:44:18.400 of two or three bricks. 00:44:18.400 --> 00:44:23.260 And that was the beginning of really radio networking. 00:44:23.260 --> 00:44:27.640 And that certainly provided efficient communications 00:44:27.640 --> 00:44:33.010 for soldiers in the field, but immediately it 00:44:33.010 --> 00:44:36.580 was recognized that that did not give you information 00:44:36.580 --> 00:44:38.560 about where the soldier was. 00:44:38.560 --> 00:44:42.370 Push-to-talk radio, the speaker gets on the net 00:44:42.370 --> 00:44:45.850 by pressing the button, and everyone else is in listen mode 00:44:45.850 --> 00:44:51.290 and hears the speaker, but we don't know where he is. 00:44:51.290 --> 00:44:54.800 All we know is that we hear his voice signal. 00:44:54.800 --> 00:44:58.600 And those radios, just to set a historical background, 00:44:58.600 --> 00:45:04.840 used analog voice over analog frequency modulation. 00:45:04.840 --> 00:45:07.090 And those of you that work in the communications field 00:45:07.090 --> 00:45:12.250 know that analog FM has a threshold characteristic. 00:45:12.250 --> 00:45:16.090 The received voice is either very good or it's very bad. 00:45:16.090 --> 00:45:18.550 And it has a threshold characteristic, 00:45:18.550 --> 00:45:20.800 and you don't know really anything 00:45:20.800 --> 00:45:24.280 about what the received signal strength is or the received 00:45:24.280 --> 00:45:26.590 signal-to-noise ratio is. 00:45:26.590 --> 00:45:28.750 So in these 50 years, we've come a long way 00:45:28.750 --> 00:45:31.240 from that primitive technology. 00:45:31.240 --> 00:45:33.850 I'm not going to go through all of this, 00:45:33.850 --> 00:45:39.670 but I want to-- halfway down, I want to mention the era of 1997 00:45:39.670 --> 00:45:46.540 when the interest began in urban and indoor geolocation. 00:45:46.540 --> 00:45:52.540 DARPA had a program that was called Small Unit Operations 00:45:52.540 --> 00:45:55.510 Situational Awareness Systems. 00:45:55.510 --> 00:45:57.580 And situational awareness basically 00:45:57.580 --> 00:46:00.160 says how do you find the warfighter that 00:46:00.160 --> 00:46:02.980 is in a hostile physical situation? 00:46:02.980 --> 00:46:06.700 How do you locate him and communicate with him? 00:46:06.700 --> 00:46:08.890 We had a piece of the research work 00:46:08.890 --> 00:46:12.280 in that project addressing specifically 00:46:12.280 --> 00:46:17.240 the radio-propagation problems in the indoor environment. 00:46:17.240 --> 00:46:22.000 And the reason I mention it is because the objectives-- 00:46:22.000 --> 00:46:24.670 the government's-- the customer's objectives for that 00:46:24.670 --> 00:46:28.550 project simply were not met. 00:46:28.550 --> 00:46:32.680 And the fundamental reason was the complexities 00:46:32.680 --> 00:46:35.830 of radio-wave propagation in the indoor environment, 00:46:35.830 --> 00:46:37.520 and that's what killed it. 00:46:37.520 --> 00:46:41.290 You could wrap all of the software 00:46:41.290 --> 00:46:45.070 that you wished around that and all the user interfaces 00:46:45.070 --> 00:46:48.880 that were all very nice, but because of the characteristics 00:46:48.880 --> 00:46:51.850 of indoor radio propagation, you could not 00:46:51.850 --> 00:46:57.070 get a precise fix on the location of a warfighter 00:46:57.070 --> 00:47:01.240 in many of those hostile-- 00:47:01.240 --> 00:47:05.140 urban fighting is the obvious scenario. 00:47:05.140 --> 00:47:07.420 About that time there began to be 00:47:07.420 --> 00:47:09.790 some commercial developments. 00:47:09.790 --> 00:47:15.610 Pinpoint evolved into another company name. 00:47:15.610 --> 00:47:17.590 I don't quite recall. 00:47:17.590 --> 00:47:22.590 Pinpoint/Wearnet came out of the body LAN technology 00:47:22.590 --> 00:47:25.830 which was also sponsored by DoD, the concept 00:47:25.830 --> 00:47:28.530 to embed sensors and communication 00:47:28.530 --> 00:47:32.730 devices into the uniforms of service people 00:47:32.730 --> 00:47:40.860 and be able to use those as part of accomplishing the mission. 00:47:40.860 --> 00:47:42.420 We'll just skip over the rest. 00:47:42.420 --> 00:47:44.070 Talk a little bit-- 00:47:44.070 --> 00:47:47.400 won't say very much about this because so much, obviously, 00:47:47.400 --> 00:47:52.380 has been said, and you folks are all well aware 00:47:52.380 --> 00:47:56.550 of concepts of asset tracking. 00:47:56.550 --> 00:48:00.030 Putting tracking golf balls in there, that was my idea. 00:48:00.030 --> 00:48:02.580 I figured I could save myself some money if I could 00:48:02.580 --> 00:48:05.340 find all those golf balls that I'm losing in the woods 00:48:05.340 --> 00:48:05.970 all the time. 00:48:05.970 --> 00:48:09.510 Actually someone does make a golf ball with a little radio 00:48:09.510 --> 00:48:14.040 transmitter in it, and I'm going to try it one of these-- 00:48:14.040 --> 00:48:19.050 but it doesn't have ID characteristics. 00:48:19.050 --> 00:48:23.430 So there's another research area. 00:48:23.430 --> 00:48:26.380 OK, try to move on. 00:48:26.380 --> 00:48:29.490 Again, I think I'm preaching to the choir 00:48:29.490 --> 00:48:33.090 here, use an old slide, but everything is a terminal today. 00:48:33.090 --> 00:48:36.840 From communications networking point of view, 00:48:36.840 --> 00:48:38.400 we don't necessarily care too much 00:48:38.400 --> 00:48:40.530 about what the device does. 00:48:40.530 --> 00:48:43.260 Either it's a terminal out at the edge of the network 00:48:43.260 --> 00:48:46.890 or it's an intermediate node somewhere within the network, 00:48:46.890 --> 00:48:56.910 and perhaps it serves both functions in some situations. 00:48:56.910 --> 00:49:00.090 This great variety of applications 00:49:00.090 --> 00:49:06.900 that keeps growing, of course, has fostered support 00:49:06.900 --> 00:49:09.720 for standardization, and other folks earlier 00:49:09.720 --> 00:49:15.090 have talked at some length about standards 00:49:15.090 --> 00:49:19.530 and the importance of standards for making an industry segment 00:49:19.530 --> 00:49:20.370 grow. 00:49:20.370 --> 00:49:26.340 And this is just our own way of characterizing 00:49:26.340 --> 00:49:31.380 some of these standards, both ultra wideband 00:49:31.380 --> 00:49:37.440 and lower-frequency technologies are 00:49:37.440 --> 00:49:41.160 being looked at for Wireless Personal Area, WPAN, 00:49:41.160 --> 00:49:45.810 one of our areas of interest, WPANs, Wireless Personal Area 00:49:45.810 --> 00:49:46.350 Networks. 00:49:46.350 --> 00:49:52.200 And, of course, the IEEE 802.11 initiative 00:49:52.200 --> 00:49:58.290 really created the renaissance for the wireless LAN industry. 00:49:58.290 --> 00:50:01.500 Obviously value in standardization 00:50:01.500 --> 00:50:07.980 and the ubiquitous use of 802.11 devices, so-called Wi-Fi-- 00:50:07.980 --> 00:50:11.880 of course, that's just a label for a certification process. 00:50:11.880 --> 00:50:17.280 But 802, the promulgation of those devices and the economies 00:50:17.280 --> 00:50:20.625 of scale that have pushed the prices down, of course, 00:50:20.625 --> 00:50:25.050 to make that an important element to be looked 00:50:25.050 --> 00:50:30.870 at in location estimation. 00:50:30.870 --> 00:50:34.980 Here again this kind of figure lots of folks 00:50:34.980 --> 00:50:42.720 use, and it just characterizes the different technologies, 00:50:42.720 --> 00:50:48.840 cellular technologies, wireless LAN and wireless PANs, 00:50:48.840 --> 00:50:55.140 against the dimensions of scale that are 00:50:55.140 --> 00:50:57.180 relevant to those technologies. 00:50:57.180 --> 00:51:05.520 Let me get to the areas of interest that have been 00:51:05.520 --> 00:51:07.560 motivating some of our work. 00:51:07.560 --> 00:51:12.960 Navigation for fireman-- a professor early in the morning 00:51:12.960 --> 00:51:15.290 had a very good example, I thought, 00:51:15.290 --> 00:51:18.150 about a hypothetical case of a fire in the building. 00:51:18.150 --> 00:51:22.340 We all have tags, and he spoke about the issue. 00:51:22.340 --> 00:51:25.560 Well, suppose there's a miscount. 00:51:25.560 --> 00:51:29.640 About seven years ago in the city 00:51:29.640 --> 00:51:31.920 of Worcester in central Massachusetts 00:51:31.920 --> 00:51:36.420 there was a very bad and deadly fire in which 00:51:36.420 --> 00:51:38.910 an unused building caught fire. 00:51:38.910 --> 00:51:43.320 And when the fire company arrived on scene, 00:51:43.320 --> 00:51:46.110 a local businessman came out and said 00:51:46.110 --> 00:51:49.980 that he saw two people running into the building. 00:51:49.980 --> 00:51:55.050 And long story short, the fire captain on the scene 00:51:55.050 --> 00:51:59.700 sent close to 20 of his firefighters into the building, 00:51:59.700 --> 00:52:04.080 and six of them got completely lost 00:52:04.080 --> 00:52:07.380 because of the smoke in the building 00:52:07.380 --> 00:52:09.240 and the structure of the building-- 00:52:09.240 --> 00:52:13.530 several floors, a number of small rooms in the building. 00:52:13.530 --> 00:52:17.760 And six of them got lost, and they died in that fire. 00:52:17.760 --> 00:52:19.740 It was a very tragic event. 00:52:19.740 --> 00:52:22.920 And after that was all over, it was 00:52:22.920 --> 00:52:25.350 discovered that the information was incorrect 00:52:25.350 --> 00:52:27.240 and there was no one in the building. 00:52:27.240 --> 00:52:31.110 And so six firefighters lost their lives 00:52:31.110 --> 00:52:35.280 putting out the fire in an empty building. 00:52:35.280 --> 00:52:40.770 We at WPI actually have now a funded project from the US 00:52:40.770 --> 00:52:44.310 government through Senator Kennedy's office, 00:52:44.310 --> 00:52:49.440 and we are looking at the use of wireless technology 00:52:49.440 --> 00:52:56.190 to try to deal with the problem of tracking firefighters 00:52:56.190 --> 00:52:59.320 in such dangerous situations. 00:52:59.320 --> 00:53:02.700 One could also call to mind the recent tragedy at the Sago 00:53:02.700 --> 00:53:08.280 Mines in West Virginia, and there is already 00:53:08.280 --> 00:53:15.570 public discussion of how various technologies, including 00:53:15.570 --> 00:53:20.820 wireless technology, might have been used to say let's move on. 00:53:20.820 --> 00:53:24.810 This is the kind of concept that the warfighters 00:53:24.810 --> 00:53:26.730 or the firefighters would like to have. 00:53:26.730 --> 00:53:28.440 They would like to have a display-- 00:53:28.440 --> 00:53:30.930 we call it a tactical display that 00:53:30.930 --> 00:53:34.530 would present some kind of representation 00:53:34.530 --> 00:53:37.380 of a building, for example, and be 00:53:37.380 --> 00:53:41.730 able to locate firefighters or warfighters 00:53:41.730 --> 00:53:45.300 within that building. 00:53:45.300 --> 00:53:47.040 This refers to small unit. 00:53:47.040 --> 00:53:53.580 I spoke about that, the situational awareness 00:53:53.580 --> 00:53:57.120 that DARPA was interested in. 00:53:57.120 --> 00:54:01.830 This I think is important. 00:54:01.830 --> 00:54:08.310 The current DoD interest is in using signals of opportunity 00:54:08.310 --> 00:54:10.530 to be able to accomplish-- 00:54:10.530 --> 00:54:13.080 to take advantage of whatever is out there 00:54:13.080 --> 00:54:16.810 in the ether, whatever frequency bands are available. 00:54:16.810 --> 00:54:25.950 And so that is a current area of interest for us as well. 00:54:25.950 --> 00:54:29.130 Other interesting research problems, 00:54:29.130 --> 00:54:34.380 location-based handoff, location-based routing, 00:54:34.380 --> 00:54:37.230 and ad hoc networks. 00:54:37.230 --> 00:54:39.840 And, of course, on that earlier list 00:54:39.840 --> 00:54:42.750 there is at least one company-- 00:54:42.750 --> 00:54:46.620 Newberry, I believe-- that is in the business 00:54:46.620 --> 00:54:52.390 of location-based authentication and security technology. 00:54:52.390 --> 00:54:56.260 There's been a lot of discussion about security issues earlier. 00:54:56.260 --> 00:55:03.810 So I'll finish up by talking about the two 00:55:03.810 --> 00:55:09.600 categories of approaches for doing location estimation, 00:55:09.600 --> 00:55:12.780 and one is received signal strength, of course, which 00:55:12.780 --> 00:55:14.650 is used in cellular networks. 00:55:14.650 --> 00:55:18.420 The second-generation CDMA networks, for example, 00:55:18.420 --> 00:55:22.170 already use received signal strength estimation. 00:55:22.170 --> 00:55:25.590 Advantage-- the hardware is simple, 00:55:25.590 --> 00:55:34.200 and that approach is not particularly sensitive. 00:55:34.200 --> 00:55:39.000 I should say the accuracy is not sensitive to the multipath 00:55:39.000 --> 00:55:40.830 and/or bandwidth. 00:55:40.830 --> 00:55:43.230 It does not require synchronization 00:55:43.230 --> 00:55:44.490 because it's incoherent. 00:55:44.490 --> 00:55:48.990 It's incoherent signal processing. 00:55:48.990 --> 00:55:54.570 However, it, in most cases, will not 00:55:54.570 --> 00:55:58.590 provide the accuracy that's required, for example, for some 00:55:58.590 --> 00:56:01.770 of these public-safety applications. 00:56:01.770 --> 00:56:07.110 Suppose, for example, you could achieve a location accuracy 00:56:07.110 --> 00:56:12.600 of one foot, and you're trying to locate a firefighter 00:56:12.600 --> 00:56:15.750 inside a smoke-filled building. 00:56:15.750 --> 00:56:19.740 Well, you may have spotted his location to within one foot, 00:56:19.740 --> 00:56:23.130 but you don't know if he's on this side of the wall 00:56:23.130 --> 00:56:25.110 or this side of the wall. 00:56:25.110 --> 00:56:27.630 So if your algorithm says he's over here 00:56:27.630 --> 00:56:30.000 and he's really over here and you 00:56:30.000 --> 00:56:33.510 have some kind of a system that supports 00:56:33.510 --> 00:56:37.500 this to try to help him find his way out of the building, 00:56:37.500 --> 00:56:39.460 he's in the wrong room. 00:56:39.460 --> 00:56:44.370 So one foot of accuracy may seem very precise, 00:56:44.370 --> 00:56:46.350 but in that kind of an application, 00:56:46.350 --> 00:56:47.700 it's not precise enough. 00:56:47.700 --> 00:56:49.005 Just an example. 00:56:51.550 --> 00:56:58.100 To achieve greater accuracy, you can resort 00:56:58.100 --> 00:57:01.350 to time-of-arrival techniques. 00:57:01.350 --> 00:57:04.560 And fundamental time-of-arrival techniques 00:57:04.560 --> 00:57:09.600 are not particularly new, but making 00:57:09.600 --> 00:57:13.290 them work in a multipath environment 00:57:13.290 --> 00:57:18.010 is a very difficult problem. 00:57:18.010 --> 00:57:23.190 An advantage is that if you can do it, 00:57:23.190 --> 00:57:26.610 you can accomplish rather accurate positioning 00:57:26.610 --> 00:57:29.220 with only a few reference points, 00:57:29.220 --> 00:57:31.260 and it also doesn't need training. 00:57:31.260 --> 00:57:36.120 The problem is that while it doesn't need training, 00:57:36.120 --> 00:57:39.090 what it does need is synchronous operation. 00:57:39.090 --> 00:57:44.160 So in communications terms, you have to build a coherent signal 00:57:44.160 --> 00:57:45.390 processing system. 00:57:45.390 --> 00:57:49.260 That adds to complexity. 00:57:49.260 --> 00:57:54.510 And you also need a synchronization process 00:57:54.510 --> 00:57:56.010 to do that. 00:57:56.010 --> 00:57:59.250 Let's move along. 00:57:59.250 --> 00:58:03.900 Just say briefly two general classes 00:58:03.900 --> 00:58:05.790 of time-of-arrival algorithms. 00:58:05.790 --> 00:58:08.400 One is distance-based localization 00:58:08.400 --> 00:58:10.140 with a few reference points. 00:58:10.140 --> 00:58:12.810 And the other, perhaps a more general way 00:58:12.810 --> 00:58:17.610 of thinking about it, is a pattern-recognition approach 00:58:17.610 --> 00:58:23.310 where you deploy many reference points on a regular grid, 00:58:23.310 --> 00:58:26.940 and then you can use a variety of pattern-recognition 00:58:26.940 --> 00:58:34.380 techniques to be able to get an accurate estimate. 00:58:34.380 --> 00:58:36.990 Getting to the end here-- 00:58:40.120 --> 00:58:46.650 a good technique in the pattern-recognition branch-- 00:58:46.650 --> 00:58:53.490 let me call it that-- is the nearest-neighbor algorithm. 00:58:53.490 --> 00:58:59.610 And I should mention Ekahau that is in the RF tag-- 00:58:59.610 --> 00:59:02.370 ID tag business. 00:59:02.370 --> 00:59:06.810 And they also have a very now highly developed 00:59:06.810 --> 00:59:13.020 software product which the Ekahau positioning algorithm, 00:59:13.020 --> 00:59:14.940 and that's very recent. 00:59:14.940 --> 00:59:19.980 And so that represents the state of the art with that approach. 00:59:19.980 --> 00:59:26.520 Our work in the center has been focused 00:59:26.520 --> 00:59:29.490 on developing an extensive laboratory test 00:59:29.490 --> 00:59:32.400 bed with instrumentation-- measurement instrumentation, 00:59:32.400 --> 00:59:36.840 channel simulation, instrumentation, 00:59:36.840 --> 00:59:40.320 and focusing on the evaluation. 00:59:40.320 --> 00:59:43.410 This gives us the capability to evaluate 00:59:43.410 --> 00:59:45.990 a variety of localisation algorithms 00:59:45.990 --> 00:59:51.780 under a wide range of realistic propagation in environments. 00:59:51.780 --> 00:59:53.660 I'll just move ahead. 00:59:56.700 --> 01:00:00.270 Coverage, of course, and range reading 01:00:00.270 --> 01:00:05.940 is a topic that's been discussed by a few speakers already. 01:00:08.640 --> 01:00:14.460 Bandwidth-- it's a common belief that increasing the bandwidth 01:00:14.460 --> 01:00:16.380 steadily increases the resolution, 01:00:16.380 --> 01:00:20.580 and therefore ultra wideband is the right solution. 01:00:20.580 --> 01:00:22.920 The problem is if you go up in bandwidth, 01:00:22.920 --> 01:00:25.830 you reduce the coverage. 01:00:25.830 --> 01:00:28.800 So there's a trade-off issue there, 01:00:28.800 --> 01:00:31.920 and we don't see ultra wideband as the optimal solution. 01:00:31.920 --> 01:00:33.960 The last topic is very important one. 01:00:33.960 --> 01:00:38.700 UDP refers to Undirected Direct Path. 01:00:38.700 --> 01:00:42.960 It sounds like an oxymoron, but you visualize 01:00:42.960 --> 01:00:44.940 a transmitter, a receiver. 01:00:44.940 --> 01:00:47.700 You're inside a building. 01:00:47.700 --> 01:00:50.730 And in many instances, the line of sight path 01:00:50.730 --> 01:00:55.290 from the transmitter to the receiver is not detectable, 01:00:55.290 --> 01:01:00.840 and all of your energy is coming from the multipath components. 01:01:00.840 --> 01:01:03.900 And what we're finding at this point that 01:01:03.900 --> 01:01:08.790 is very often the Achilles heel for time of arrival based 01:01:08.790 --> 01:01:11.670 positioning estimation system. 01:01:11.670 --> 01:01:15.540 And that, in fact, was the central problem 01:01:15.540 --> 01:01:18.840 that caused the failure to meet the objectives 01:01:18.840 --> 01:01:25.960 in the [INAUDIBLE] program several years back. 01:01:25.960 --> 01:01:31.140 So we continue to focus our research now 01:01:31.140 --> 01:01:35.760 on algorithms that will allow us to operate 01:01:35.760 --> 01:01:38.880 in a condition of undetected direct path, 01:01:38.880 --> 01:01:41.940 and that very often means making use 01:01:41.940 --> 01:01:44.940 most of the time of the multipath components. 01:01:44.940 --> 01:01:49.590 And we're looking at techniques like tracking, which works fine 01:01:49.590 --> 01:01:51.960 if the transmitter is mobile. 01:01:51.960 --> 01:01:56.190 However, that doesn't work if the transmitter is not mobile. 01:01:56.190 --> 01:01:58.330 If it's mobile, you can do tracking, 01:01:58.330 --> 01:02:01.290 and you can work with the multipath components 01:02:01.290 --> 01:02:06.810 and the direct path, which will occasionally appear in a time 01:02:06.810 --> 01:02:09.336 record of measurements. 01:02:11.880 --> 01:02:19.170 We're also looking at use of diversity techniques. 01:02:19.170 --> 01:02:23.700 And some of that is being done in cooperation with Draper Lab. 01:02:23.700 --> 01:02:25.620 And I realize we're running out of time. 01:02:28.500 --> 01:02:34.350 So just beating the drum and saying that localization 01:02:34.350 --> 01:02:37.020 is still an important research area, 01:02:37.020 --> 01:02:39.180 and we regard it as an unsolved problem. 01:02:39.180 --> 01:02:40.410 Thank you very much. 01:02:40.410 --> 01:02:42.260 [APPLAUSE]