 the webinar series of the IT Journal on Future Involving Technologies. My name is Alessia Magliarditti from ITU, the International Telecommunication Union. ITU is the United Nations Specialized Agency for Information and Communication Technologies. ITU allocates frequencies to the services that make use of the radio communication spectrum, it develops standards and assists developing countries in setting up their information and communication infrastructure. ITU and academia share a commitment to the public interest and this commitment is embodied by the IT Journal which offers a complete coverage of communications and working paradigms free of charge for both readers and others. Our journal welcomes emissions at any time on any topic within its scope and we believe that this new webinar series will inspire more contributions from researchers around the world. It is my pleasure to open this webinar today with Professor Joseph Jordanette from Northeastern University from the USA. We count on your support to make this webinar an interesting experience. Please submit your questions via the Q&A channel. We will address them to our speaker during the Q&A session. After the talk and the Q&A, please stay online. We have something very special for you, the wisdom corner, live life lessons. Professor Jordanette agreed to a very personal chapter. He will share with us some lessons learned over the years that might perhaps be useful for some of you. It is now my pleasure to introduce Professor Iana Kilditz, Editor-in-Chief of the IT Journal and President and Founder of TRUBA from the United States. Almost two years ago with Professor Iana Kilditz, we established this new scientific journal and after almost two years we are probably moving towards impact factor. Professor Iana Kilditz is Ken Bayer's Chair Professor in Telecommunication Emeritus at the Georgia Institute of Technology. In the last two decades, he established many research centers worldwide, including in South Africa, in Finland, in Spain and Saudi Arabia. He is Editor-in-Chief Emeritus of Impact Factor Journals and highly cited and at the top of the most prestigious international rankings. He is visiting distinguished professors in several universities around the world. And his current research interests include the 67G wireless communication systems, hologram communication, Therahertz communication, Internet of Bio-Nano things, intelligent surfaces, molecular communications and many other subjects. So, Professor Kilditz, the floor is yours for your opening remarks and to introduce our speaker. Thank you. Thanks a lot, Alessia. Good morning, good afternoon and good evening and good night, wherever you are in the world. Thanks for joining us. Today we have Professor Joseph Jornet as our speaker. He is the last speaker of the season. He will talk about Therahertz band communication and I would like to introduce him instead of reading his biography because I have a personal connection to Joseph. It's like almost like my son. So, I met Joseph in 2008 in the summer at University of Polytechnic in Catalonia when I was working with them. We started this nano networking center in Catalonia and his professors there told me he's a genius at the UPC and I should try to hire him to my lab and I really got scared. Why? Because I do not want to lose him. I told them, don't pressure him, please. Otherwise, it's like a little bird, right? It will fly away and plus, he was doing his master's degree that time with MIT and so when I met him, I told him that I'm really interested in these nano networking, like nano scale devices and communications and he was cool in the beginning, which is good actually and then I was also teaching cognitive radio networks that time and he came to my classes and he was always asking interesting questions and I was hoping that he would join my lab again as I told you it was summer 2008 and then he came to me I said, you know, I really like your ideas about nano scale machines and networking and he wanted to know more so we started to talk about it and then he said he wants to do something brand new. That's why he's really inclined to work on this topic and then he didn't tell me anything, right? So that was another very nice move. So like in September, October, I guess, he said, I'm coming. So I was really very happy that he decided to join my lab and then guess what December 31, 2008, meaning January 1, 2009, he came to Atlanta and then we started to work. It was an incredible greatest pleasure that I had. I produced like almost 50 PhDs and I always use this word politically correctness. He was one of my top students but I can say that he was the top in my opinion and not only like, you know, research-wise but also, you know, we had great time together like any topics. He's like a jewel of a person really and so we did a lot, you know, we made it, achieved a lot of success. Like I really am proud of these things. Like we moved nano-networking like then Terahertz, this was like 2009-10, right? So then we were of course pondered by some people like, you know, there's a science fiction and all that and we received three super patents like nano-antenna and nano-transceivers and also we called this ultramassin limo and then he graduated in 2013. It was time for him to fly, right? I mean, I would have kept him 10 years but I'm against those people that they do that to their students. So then he started his career. Of course, I was very hopeful and then he delivered what I had hoped and he continued his research and he published and he's still publishing a lot of papers. He got received many, many projects and funding as well as many awards and he took over the ACM nano-com conference and then also the Elsevier nano-networks journal and he's leading them very successfully. He has many, many students in his lab working on forefront and I remember we used to discuss like again, like 14 years ago, this is really brand new and we are getting pondered. I used to tell him just be patient and you will see that the impacts will show up and guess what? The last three years really into our world is working on Terahertz as if it's brand new thing, right? So it's really funny. So anyhow, I'm extremely proud that Josep invited, accepted our invitation and I want to say one more thing because this is my philosophy in life. I heard that also through him and others, you know, what is the role of the advisor, right? So let me explain that to you also for the world, hopefully this will go out. So I see the academic career same like in other fields, like in sports, like football, like soccer or Hollywood movies. I want to be in Hollywood but anyhow, I couldn't achieve it. So what I, what I say that is, so if you have an excellent trainer, right? Like, you know, Guardiola for example, for me is really top and then he has the excellent players in Manchester City or Jurgen Klopp in Liverpool. But these people can only be successful when they also have super players. Same thing in the Hollywood, right? When the director is excellent and the actors are excellent, outstanding products come out. And same thing here. I was extremely lucky or whatever you call it. I found people like Josep. I found these people, of course, through my friends. And we had great time. It was not easy. I'm sure you can tell it when it starts to, but we achieved fantastic things. So in other words, you have to have a good advisor, an excellent student and excellent products come out. Okay. So again, thanks for listening to me, Josep. It's your turn. Thank you again. Thank you very much, Professor. I mean, I don't know how I'm going to give the talk now after this overwhelming introduction, but I'm sure that when we get to the end to the life, life lessons, we will have many good things to mention. But thank you again for the invitation and thank you for enabling me to do the work that I'm going to present today. So with any other further delay, let's just start the conversation. Remember, you can write questions in the Q&A and then at the end we'll have time for those and for some other questions in the back. With that, we're going to be talking about Thetahertz communications, but not in the same way that we talked about this five years ago or 10 years ago. We're actually going to be showing solutions that work and that we already have today to make the most out of the Thetahertz band. I'm currently a professor at North Eastern University. I direct a very exciting group of students from everywhere following how I was trained in a very exciting interdisciplinary lab. And what I'm going to show is the result of not just me, but of a team. So let's do this. Okay, I think that at this point, everyone on this call knows that, well, what keeps the academic and the industrial community running these days are the discussions on 6G. There are many metrics that are being shown about what 6G should do. One of them that keeps coming and coming, and it seems that both academia and industry seem to agree, it's on the target data rate. And it seems that one Theta bit per second, it's finally going to be something that we want to be able to achieve practically. Now the question is, how do we get to one Theta bit per second? Well, it depends on what type of modulation do we use, but also if we can exploit several parallel channels. So this is what we show exactly in this figure. In this figure we show the bandwidth, that's the height of the columns as a function of the modulation order. And we start with 16, like a 16 one, but we go all the way up to 1024 one, just for example, and the number of parallel channels. We don't even show what channel because we know that we're going to need a few of them, but let's be conservative from two up to 64. What we see here is that unless we are really able to work with this very high order modulation 1024 and large number of parallel channels 64, unless we're able to do that, we're going to need bandwidth, which is much more than what we have today. For example, in 5G we're dealing with bandwidth, which are up to 400 megahertz and actually you can aggregate two of those channels and get to 800 megahertz of bandwidth. Well, if you want to go to one Theta bit per second, you will have to get much more than that. You will easily need 10, if not more, gigahertz of bandwidth. Now the question is, where do we find such bandwidth? And well, as you all know, this is a frequency axis that starts at zero and goes all the way up to 10 Thetahertz. If you think of the spectrum under 100 gigahertz, pretty much everything has already, I mean everything has been allocated and pretty much everything is already being used. The result of that is that we don't have contiguous bandwidth that can be used for these very high data rates. Of course, there are always things you can play like, you know, channel aggregation, this and that, but we're not getting close to what we need. Of course, you may say, well, if the only thing that we need is more bandwidth, we could jump to the other end of the spectrum and get to optical communications. And that's fine. That's a very exciting field. People have been working on it for many years. But the question is, why would you go there? I mean, why would you skip? What happens between the 100 gigahertz and the much higher frequencies, hundreds of Thetahertz, at which optical systems operate? Well, what's in between? It's the Thetahertz band. And this is the area in which we have been working since 2009. As Professor Achilles mentioned, you know, when we started working on Thetahertz communication, and I show it in the next few slides, we were thinking that Thetahertz communications were good for the short range. And that's what we were looking at. Then we realized that actually we can do better. We can do also longer distances. And that's fine. More recently, you know, everyone has started to jump in. So literally last week or few weeks ago, we had this new position paper publish in which we actually explained what has already been done in Thetahertz communications and what still needs to be done. In other words, Thetahertz communications, it's not a new field. I mean, it's exciting and it's new. There are many new problems to address. But it's not something that we just woke up yesterday and said. So with this in mind, let's go and see what is there for Thetahertz communications. And we'll start talking first about the applications. We are in the context of a wireless communication presentation. So we know that the presentations, I mean, the applications of the Thetahertz band are going to be in communications. We can think of some traditional applications, like for example, try to get you those Theta bit per second links in short range, like in wireless personality networks or wireless locality networks. The reality is that through the years, the technology has improved, our understanding of the channel has improved. So today, we are comfortable talking about, for example, creating long distance wireless back holes, also able to carry Theta bits per second link at Thetahertz frequencies. And you know what utilizing that same, the same tools, that same knowledge that allowed us to predict that we can go for short to long to longer range, we are comfortable saying that we will be able to use Thetahertz communications, also in the context of a space networks. And we'll talk a little bit about that. Of course, communications are very exciting. And what I want to say is that, look, one of the things that usually happens in a new generation of systems is that, well, the first one that gets access to the other people who can afford it. But for example, through the Thetahertz wireless back hole, we might be able to have real implications beyond just the rich people who can buy the most expensive phones. We can actually try to bridge the digital divide by bringing broadband or ultra broadband connectivity to everywhere, including the rural areas where optical fiber doesn't make sense. For example, this is half of the story, however, because the reality is that the Thetahertz band can also be utilized for sensing. In fact, in 2009, when we started working on this, the only papers that we would find relating to Thetahertz technologies or the majority of the papers related to Thetahertz sensing. What does sensing mean? Well, sensing means different things for different people. For example, sensing might mean radar or localization. And we know that when we go to higher frequencies with a smaller wavelength and larger bandwidth, we can see smaller things. And that's great. We can have more precise, higher resolution radar. What's interesting is that Thetahertz radiation can also interact with the atoms, with the molecules of the materials. And we can use that to not just know where the things are, but actually understand what are those things composed of. So we can use non-damaging imaging and spectroscopy to see the context of a box without opening the box and without damaging the context of the box, as opposed to X-rays. What's also interesting is that right now, as we speak, there are already satellites orbiting the Earth with Thetahertz radios in them. Those satellites are not being used for communications. Those satellites are in fact characterizing the atmosphere. And yes, the technology was ready several years ago when those satellites were launched. And for example, that technology is what's being utilized today to, for example, understand the climate change and try to study those. Now, communications and sensing, as you know, do not need to happen separately. They overlap. And that's another very exciting application. For example, let's put all these things together. So we can think of, let's say, an indoor scenario where we have people in a meeting room, people presenting, people actually with the virtual reality heads that may be talking to someone in the metaverse, all those things are going to happen. Now, where does Thetahertz enter the game here? Well, the Thetahertz band on the one hand, it's going to give us those very high-speed links that will allow you to experience a 4K or 8K 360-degree full immersive virtual reality. Or at the same time, do any other type of high-data rate-demanding application. Of course, we will learn that Thetahertz radiation sometimes does not propagate as well as we would like it. So there are going to be blockages and things like that. So we expect that intelligent, reflecting surfaces or whatever you want to call them are going to enter the game, mostly to overcome the lack of line-of-sight paths. This is great. But what's interesting is that at the same time, when all this is happening, we can be doing sensing. And through that sensing, we can, for example, prevent the person with the virtual reality headset to actually bump into something or into someone. And at the same time, just because of how Thetahertz radiation interacts with the air we breathe, we can even, for example, extract the quality of the air we are breathing. This is not science fiction and this is not just a fancy picture. This is actually what Thetahertz radiation allows us to do. And we can do this indoors or we can go all the way to space if you want. As I mentioned, there are already satellites orbiting the earth with Thetahertz radios. And actually, they are not really transmitting. They are listening. They are collecting signals both from the atmosphere, as well as from the space trying to characterize the origin of the universe and the quality of the air we breathe. So in those satellites, we want to create communications. And that's great, point-to-point links. We could also create distributed MIMO systems in space. And I know that, again, this sounds more like a James Bond movie, but we have the technology to get there. And at the same time, we can be doing that while collecting exciting atmospheric information. And for example, we just, again, it sounds very good when you have a PowerPoint picture, but we just show you how we could do that by being smart at the physical layer. Okay, what's interesting is that you think, okay, so this is the Thetahertz band. Let's go. No, what's interesting is that this is only half of the story, because all these applications that we have just shown you are what we call macro-scale applications, applications in which at the end of the day, we are using a device that we can touch with our hands, that we can see. And this device is the one that communicates with the network. But as Professor Achilles mentioned, when we started working on Thetahertz communication, we were actually not looking for a macro-scale solution. We were looking for a nano-scale solution. We were actually aim at creating communication links between devices that are as small as, for example, the cells in your bodies. I'm talking at the very most cubic micrometers, ideally less. And what are the applications of those type of devices? Well, we can talk about wireless networks and chip. And for example, our colleagues in Barcelona in the nano-networking center that Professor Achilles funded are heavily leading, I would say, the field of wireless networks on chip at Thetahertz frequencies. We can talk about the Internet of Nano-Things or Nano-Bio-Things, or in general, wireless nano-bio communication networks. And we can do that not just for communications, but also for sensing. Again, today, when we mention this, you say, okay, yeah, this is kind of doable, right? Because it doesn't sound too crazy. Now, go back in time. This paper, the Internet of Nano-Things, was a paper that we created in, I mean, it was published in 2010, which means that it happens around 2009 and beginning of 2010. And, you know, to write that type of paper, well, what you need, it's a visionary person. And it was not me. It was someone who said, let's look into this. That's again, Professor Achilles. Now, let me show you how, for example, this joint nano-bio communication and sensing could look like. If you go to Google right now, and you look for nano-biosensors, you will easily find, let's say, 100,000 papers. And I'm not exaggerating. I would say that easily 60 to 70% of these papers deal with Plasmonic nano-biosensors, which are devices that at the end of the day, you engineer them. You add some chemical layer in such a way that when you find some specific biomarkers, for example, in your blood, these chemical properties change, that means that the actual reflection properties of the chip change, which means that then you can just do radar, if you want to imagine it with this chip, to know what's going on. You could do that in a lab, but instead of doing that, we are communications people. So what do we want to do that? Well, we want to interrogate this device when it's implanted inside the body. And to do that, for example, we want to take advantage of wearable devices. And of course, the goal of the wearable device is not just to excite and collect measurements, but then to share it eventually over your phone or to the internet to your healthcare provider. And again, these two days, you say, sure, implantable, yes, something under the skin makes sense. Sure, a wearable device makes sense. Of course, everything is connected to the internet, but now traveling time 12 years. Then you tell me if it sounds crazy or not in 2010. Now, hopefully at this point, you're excited and you say, okay, these are very interesting things. These are things that we totally want to work on. Well, and that's great. There are only some challenges that we have to take into account. Some of these have been solved in the last few years. Some of them are still pending. And what we're going to do in the remaining of this talk, first of all, go through this pyramid. So we'll start talking about what type of materials and devices we can utilize for the rehearse communications. What happens with the channel? Is it as bad as people seem to think? Once you know the devices and you know the channel, what type of signals should you use? What type of communications and signal processing should you be doing? And remember, we're trying to not just make a point-to-point link. We want to make 6G or 7G networks. So we will have to work on networking. We are in an ITU presentation. So here, more than ever, we actually need to highlight the importance of spectrum policy and regulation, which at the end of the day, unless that exists, we cannot use any of our technologies in the open. And we'll talk a little bit about a standardization tool. I know there seem to be many boxes. You will see that we'll spend more time on some than the others, but I promise that everything will be connected by the end. Let's start with the bottom. Let's talk about materials and devices. Okay, once again, travel back 12 years, 2009, 2010. Every single work on rehearse communications would start. There is a technology gap when it comes to the Terahertz band. We call the Terahertz gap or the Terahertz technology gap. Do we have a technology gap today in 2022? I would say there might be a gap, but actually, we're almost done. I'm not saying that everything is done. I'm just saying that we certainly see the light and we're going to show you some devices that work at Terahertz frequencies today. First of all, what do I mean by Terahertz technology? Well, if we had to build a Terahertz radio, what do we need? First, we need the analog front ends. What do I mean by front ends? The front ends means that the devices that can generate, modulate, amplify, filter the signals that ideally should be at Terahertz frequencies. And throughout this presentation, we use Terahertz in the broader of its definition. So from 100 gigahertz all the way up to 10 Terahertz. So we want to make devices that can operate at these frequencies. And of course, we want these devices to also have large bandwidth. Remember, as much as we love the Terahertz band, it's not just a matter, all it's just Terahertz frequencies. No, if you're thinking of communication, you want to use the Terahertz band because of the large bandwidth. If you cannot leverage the bandwidth, or if you don't need the bandwidth, maybe you don't need the Terahertz band. Let's be clear. Let's not exaggeratively hype the field. Now, when it comes to the Terahertz band, I would say that traditionally, there were two paths that called the electronics path and the photonics path. The electronics path, it's what we know as up conversion. You generate a signal as you do in microwave or millimeter wave signals. And then through frequency multipliers, you take it up in frequency to the Terahertz band. Then there is a photonics case, which is the other way around. You start at optical frequencies, hundreds of Terahertz, and then through, for example, laser multipliers or photo conductive antennas, you take that signal down. That's called down conversion. There is a third path that actually we created, I would say, with Professor Achilles, which is what we call the plasmonic path. And I will not show it to you just yet, because compared to the others, it's a little bit behind. But I promise you, I'll show you at the right time. Now, if you want to buy today a Terahertz radio, can you? Yes, you can. And for example, this is a picture of some of the electronics based radios that we have in our group. Some of these come from commercial vendors like Virginia diodes. Well, now it's not a spin off anymore. It's established company out of the University of Virginia. Some of them comes from NASA, JPL, the same team that puts the radios on satellites. We can collaborate with them and have access to them. And for example, to give you an idea, these are radios that are on the lower end of the Terahertz band. So few hundreds gigahertz with powers that are not too bad. For example, 200 milliwatts, which by the way, it's close to the world record, or it used to be world record. It's very high power at these frequencies. Okay, that's just one building block of your transmitter, which is, well, the device that can at the transmitter, generate, modulate, amplify and filter and at the receiver should be able to recover. Fine. What's the other element that we need? Well, we need antennas. We need antennas that can convert these electric signals or optical signals into an electromagnetic signal that propagates. And we need to be able to manipulate these Terahertz signals, both at the transmitter and the receiver. And well, in today's society, we understand that we also need to be able to do that in reflection. Right? So again, we'll talk later on about intelligent reflecting surfaces. That's a given. We know that we, if we want to create intelligent propagation environments, we need the transmitter, we need the receiver, and we need the channel to help us where the channel contains smart surfaces. Now, we will see that we will need high directivity gain. And that's something that we'll have to deal with. And another thing that we would like to say is that, well, the Terahertz band is between microwave, millimeter wave and optics. In optics, usually we don't use antennas. We use lenses or even programmable lenses, which are made on metasurfaces. Well, at Terahertz frequencies, we can also learn from those. And for example, many of the antennas we have are a combination of an antenna and a lens. The antenna will help you radiate in a given direction. The lens will help you focus. Lenses traditional were just, you know, a piece of plastic, a dielectric lens that you print it or you design. But today, thanks to metasurfaces, which are nothing but flat metamaterials that can be tuned, we can create programmable lenses. And to give you an idea of the order of magnitude, I would say antennas with 20, 25 dbi of gain are very common, but even higher gains, like 40 or 55 dbi of gain per side. So at the transmitter and then at the receiver are common and some of them even commercially available. That's fine. So it seems that we're done. Are we done with the Terahertz radio? Actually not. There is one more thing, which is that, well, we're going to the Terahertz band mostly because of the large bandwidth. So what do you need? Well, you need something that can digitally generate your signal. I mean that that can pump the zeros and ones as fast as your radio can swallow. And that can then take the zeros and the signals as it comes. So you can recover the information at the receiver as fast as your Terahertz radio allows you. So well, the digital back end, that's what we call the interface between your computer, which is a digital machine and this analog systems. And this is a very important element. Of course, how do you get to there? How well, you may remember from your very basic signal processing class, right? If you want to operate with signals that have a given bandwidth, you need a sampling frequency that it should be at least twice that bandwidth. That's the Nyquist theorem, right? So if you want to operate with a signal that I don't know has 50 gigahertz of bandwidth or 100 gigahertz of bandwidth, you will need digital to analog and analog to digital converters that can sample at least twice that bandwidth. And that's challenging. What's happening instead? Well, or in parallel, well, one option is, well, don't think of this big bandwidth as just one big band. Think of it as a collection of narrower channels and then try to take advantage of highly parallelized systems. And for example, one of the platform that is becoming very popular are these radio frequency systems on a chip, which are think of it as an FPGA with a bunch of analog to digital and digital to analog converters. Each one of them can only handle two, three, four gigahertz of bandwidth, but you can multiplex them and then together you can attack this much larger bandwidth. So I'm showing you pictures of things that work and I took just the things that work in our group. There might be others. But what I want to tell you is that you cannot start a paper in 2022 saying that there is a technology gap and that's why you only are going to do some very high-level theoretical assumptions. I think that we are at a point that we need to take into account the reality and the reality is that at least in the lower end of the Terahertz band, we have some solutions. Something similar happens when we talk about the channel, which is like there are still, you know, some newcomers to the field that we say, oh, we don't know how the Terahertz channel is. We need to do all these models. We need to do all this work. And the answer is that, well, actually we know how the Terahertz band channel looks quite well. We know that the Terahertz frequencies, you have access to very large transmission bandwidth and we know that this comes at the cost of a very high propagation loss. First of all, there is molecular absorption. What does molecular absorption mean? Well, molecular absorption means that think of electromagnetic radiation. Once again, go to the basics, Terahertz frequencies, Terahertz radiation. Remember that when you talk about an electromagnetic radiation, you can understand it from the wave perspective or from the particle perspective. If you think of Terahertz as an extension of microwave, a millimeter wave, think of it as wave. But for some things, it might be better to think of Terahertz as low frequency optical systems. And low frequency means low photon energy. What's interesting is that Terahertz photons can provide enough energy to molecules, particularly water vapor molecules, a little bit less but oxygen too. And those molecules will start shaking. And you know that if those molecules are shaking, they have kinetic energy. If there is kinetic energy, and given that energy cannot be created or destroyed by only converted from one way to another, if suddenly molecules are shaking, it means that you have less energy. What's interesting and what's very important is however that this molecular absorption only comes at some frequencies, at the frequencies at which your photons have the right energy for them to be absorbed. And that's why in this figure, in which we are plotting the loss as a function of frequency, you can see focus on the solid lines. So you can see that the solid lines have these spikes. These spikes are the molecular absorption peaks. That's where molecular absorption happens, which means that yes, there is absorption of Terahertz frequencies. But there are also many frequencies within the Terahertz band that do not have absorption, or where the absorption is maybe five, six, seven dBs. And you know what? We can deal with that. We can deal with five, six, seven dBs. We probably don't want to communicate on top of an absorption line with 150 dB of loss, but we want to communicate in places where there are just a few dB of losses. Of course, of course, if you're trying to build maybe a satellite communication system and you don't want that anyone eavesdrop you from the ground, maybe then you want to design your system on top of an absorption line, because above the atmosphere, no one is going to hear you. In any case, to overcome molecular absorption, we just need to pick the right signals, pick the right frequencies, pick the right bandwidth, and be intelligent with it. And we'll talk about that later. Of course, does it mean that the Terahertz band channel is great? Well, not really. The problem is that we do have very high spreading losses. Those are the dotted lines. The spreading losses, we call them that it's a channel problem, but in reality, it's an antenna problem. What happens is that when you go to higher frequencies, your antennas become smaller, and a small antenna can intercept a smaller fraction of power. So even if your emitter is sending as much power as you want, your receiver can only intercept a tiny fraction of the power. So what should you do? Well, instead of just sending the power everywhere, you should send the power in the right direction. So what do you need? You need directional gain. You need to focus your signal in space. There is another problem, which is blockage. We said that Terahertz radiation interacts with obstacles, and for example, that's what enables sensing. But again, they will not go through many obstacles. So we have blockage. And what will we have to do? Well, we'll need, for example, intelligent reflecting surfaces. At low frequencies, sometimes some people are not convinced about the need of smart surfaces. Well, at Terahertz frequencies, that's what you may need if you want to get rid of blockage. Now, both to overcome the spreading losses and the blockage, what you need is directional beams that you should be able to point in different directions. And well, to do those directional beams that you should be able to steer, what do you need? You need antenna arrays. If you see, I didn't cover that in the bottom box, but I'm covering now. So you need antenna arrays. Are there antenna arrays at Terahertz frequencies? Well, there are some arrays in the lower end. Like between 100 and 300 gigahertz, there have been some results. Like, for example, Kaushik Tsengupta, who is now at Princeton, when he was a PZ2, he was already looking into non-traditional ways of creating arrays at these frequencies. But for example, if you look at the work that's coming out from the SRC communications and sensing at Terahertz frequencies center out of UC Santa Barbara and a bunch of others, you can see that they have been building arrays at around 140 gigahertz with reasonable power, reasonable bandwidth. The only thing is that these arrays have, well, between eight and 16 parallel channels. And when you ask them, could you increase that? Could we go to higher order of parallel channels? Or maybe can you give me more power? How much can we do with these arrays? Well, the answer is that, well, when you go with these traditional technologies that are mostly CMOS, the main challenge that you have is that, well, the antennas become very small when you go up in frequency. But everything else around the antenna, your signal generation, the generators, your mixers, your amplifiers, your filters, do not scale that quickly. So what's happening is that if you want to go with this, I would say traditional approach, what happens is that it's difficult to package everything in a chip, mostly because of photo thermal problems. Okay, well, what can we do? Well, just thinking the following way. So far, we have kind of shown you example of what's known as the electronic and the photonic approaches. And in both cases, we're dealing with systems that you start in a different frequency and then you try to go to the right frequency. So you're doing frequency converting systems, either from the bottom up or from the top down, you're doing frequency converting. The reality is that every time that you do one of these conversions, first, you need the special devices, which are not easy. But also what's happening is that, increasingly, you're actually generating harmonics and other mechanisms through which you lose energy. So the question here is, could we create devices that intrinsically operate in the terahertz band? Can we make something very simple that when you connect a battery to it, it starts radiating a terahertz signal without the need of going to microwave or optics? Could we do that? And the answer is like, well, we think we can. And actually, this is interesting. We start working on this technology back in 2009 and then with Professor Achilles, this is what we call the plasmonic approach, particularly enabled by graphene. What am I talking about? Well, in this, we actually, I'm going to show you three blocks. Let me put them all on the screen. And we're going to do the following. Think in the following way. We're going to start from the left. On the left, what you have, it's a transceiver. What is a transceiver? It's a device that should be able to generate a terahertz signal or reciprocally detect a terahertz signal. Well, what we propose is that, actually, it was already in the 19th that someone explained that you can take a very small transistor, a high electromobility transistor. And if you have the right material and you create the right boundary conditions at the source and at the drain of this transistor, you can actually excite terahertz plasma wave. Terahertz plasma waves that are oscillations of electrons at terahertz frequencies that will be confined in the channel of this device. This device, to give you an idea, it's around hundreds of nanometers long. The problem is that in the original design, that signal was trapped under the gate of this transistor. Instead, what we realize is that, well, there is a material, a material that in 2009 and 2010 was a very new material. I'm talking about graphene, the first two-dimensional material. Now, there are many others that if you want, it can look like a metal. So it can be used as a gate for the transistor, but it is not really a metal. It's a semiconductor. So actually, it can support the propagation of this plasma wave, which no longer will be confined in the channel of the device, but they will be gliding on top of the graphene layer. Well, we actually show that this could work and we've got the pattern for this. So we show that there is a device that it's hundreds of nanometers long that at room temperature can lead to the generation and radiation of terahertz signals. Of course, I'm talking about a device that is hundreds of nanometers long. So do not expect much power out of this. It's not that it's not efficient. It's that it's small. But okay, that's one thing. We have a device that can generate signal. The second device is that, well, I need to add information to it. So I need to be able to change something on this signal. And in this case, we can take advantage again of graphene as a material that by applying a voltage to it, I can change its properties. And I can change, for example, the speed at which your plasma wave will propagate from left to right. So if I can change the speed, it means that I can control the delay. It means that I can control the phase. So we actually created a phase modulator. And on the right, you have the last device that you need, which is a device that allows you to radiate the signal out. This is a signal that it's on graphene and what you get to get it, it's on the air. So what do you need? You need an antenna, but an antenna that converts these plasmonic waves into free space electromagnetic wave. What's interesting, what I would say is that if you look at the date of these works, we actually started with the antenna. Our first thought of how do you utilize graphene was in 2009 and published in 2010, which is how to use graphene to make these very small antennas, antennas that we could incorporate in nano devices. When we publish this work, I'll share the experience later at the end, you know, it's okay. It took some time for this idea to fly off. But then what we realized is that while in addition of the antenna, you need to be able to excite the antenna. So we went to the transceiver and then of course you need to add information. Fine. These devices are great, but these devices are very small. I'm talking about devices which in total are a few micrometers. So no, these devices are not going to give you buds of teraherspower, but these devices are very small. So you can place them in your nano biosensors. You can place them in your nano networks on chip. But what's also interesting is that they are so small and not just the antennas, everything else, that you can actually place them in very large arrays. And it was again, 2016 or a little bit before that because as you know, papers take some time to go down. So around 2015, 2016, again with Professor Achilles, we said, you know what, we have very small antennas and it's great for our nano scale applications. But the big opportunity for terahers is not just nano, it's also macro. So we need to leverage this technology for macro. So well, idea, well, we can put many of these elements on a surface. We can build very large arrays with this. And what's interesting is that each element, it's not just one antenna. Each element has its own source, it has its own modulator. So actually each element, it's a full front end, if you want to think like that. What's also interesting is that these elements are smaller than the wavelength. Again, being smaller than the wavelength means that maybe they are not the most efficient radiators. That's okay, we can discuss about that. But you can put many of them very close. And we know how close we can put them without getting into the mutual coupling region. So we're fine. So we can make very dense arrays with many elements, which are sampling space. Just think how an array work. An array is nothing but a device that samples the wave as it comes in space or emits a wave that goes from your digital conception of the wave to an analog wall. So we can sample space with much more resolution than your typical have a wavelength. What does it mean that we can do very interesting things? This is what we call ultramassive MIMO. This is what we actually got a pattern for. And what I would like to say, no ultramassive MIMO, it's not just massive MIMO with more antennas. It's taking advantage that these antennas are closer than have a wavelength. And therefore, with these antennas, we can do things that with traditional technologies would not be able to do. I'll show them in a couple of slides. You'll see where I'm getting. Now, what's interesting is that this ultramassive MIMO can work in transmission, can work in reception, and can also work in reflection. There is nothing that prevents us from utilizing this technology, for example, in reflector rays. We would call them plasmonic reflector rays because our elements are less than have a wavelength. We could even call them a metasurface, but I would not try to mix the things. But the bottom line is that you can use them in transmission, reception, and reflection. And again, we have some follow-up works on how this could be practically implemented. This is work in progress, but it's happening. Talking about the reflections, that's what I wanted to get. Look, it also seems it's very interesting. I think it was a couple of years ago or three years ago that there were this group in a given country claiming that the original paper on intelligent reflecting surfaces from our group gets this so many citations. And I'm like the original paper on intelligent reflecting surfaces. First of all, intelligent reflecting surfaces is not something new. They have existed for many years, actually, from decades. The problem is that maybe we didn't have a real application from them, but just always be careful when making some of these claims, right? Because we'll get to that later. But with so many papers being published today, it's easy that some good papers are missed. And maybe there are good papers being published today that they will not pick till 10 years down the line. Do not forget about those papers that are happening today that in 10 years will have merit. But we'll talk about that at the end. Now, what I want to say is that there are intelligent reflecting surfaces at Therahertz frequencies. Or actually, there are reflecting surfaces at Therahertz frequencies from several years back. Usually, these are done by groups that come from the optical domain. So they are used to do with small structures. The main challenge that we have when we go to Therahertz frequencies is that we can make reflectors that can operate in a given way, but it's difficult to make these reflectors tunable. By tunable, I mean that you would like to be able to engineer in which side your signal is reflected, which means that you need to have a device that it's controllable at Therahertz frequencies. This has been hard, but you know what? Well, that's exactly what graphene is. Graphene, at some point, like in 2010 and on, it was, there was a lot of hype about graphene. Graphene was supposed to be the material for everything. And the reality is that there have been many applications that did not really take off. But when it comes to the analog applications of graphene and it's used in high frequency applications, it actually makes sense, because graphene is a very good conductor at Therahertz frequencies. That's why we can make the sources, the modulators, and antennas, but it's also tunable. So we can use that as the tunable element in your, in your reflector ray. And for example, that's what we did. As I said, graphene, it's a semi-metal. So sometimes it looks like a good metal, but it's not really a good metal. Sometimes it looks a semiconductor. So what, for example, we've done with Arjun, one of my students is, do you know what? Let's don't go only with graphene. Let's use metal as a reflector element, but then use graphene to control the reflection phase of this element. With that, you can actually go and design systems that can give you the tunability that you expect from an intelligent reflecting surface. But at the same time, by having a hybrid element, we can actually be more efficient if we're only graphene. So the problem is, as I said, today it's difficult to make metallic tunable reflecting surfaces. Why? Because what's the tunable element? Fine. We can use graphene, but if we use graphene only, you are tunable, but then you're not too efficient, you can combine the two. You can make hybrid designs. Fine. At this point, let's just stop for a second. Remember what we said, we said that when you go to Terahertz frequencies, your wavelength becomes smaller. Therefore, your antennas are smaller because your antennas usually go with half a wavelength. And we said, well, because our antennas are smaller, our effective area, it's small. We cannot intercept much power. So what we need to do is to, again, go to directional antennas. We need to make them larger. We need to go, for example, to a race. Fine. So that's okay. We have antennas that can be very small, but we put them in large structures. What happens is when you make a large radiating structure, large in terms of wavelength, it's still a small. I'm talking about when I say large at Terahertz frequencies, I mean centimeters or maybe less than a meter because the wavelength is so much smaller, right? We're talking about wavelength, which are less than a millimeter. What happens is that when you make these larger structures, whether it's your ultramassive MIMO, the transmitter at the receiver, when you make a reflecting surface at Terahertz frequencies, what happens is that your far field. So the distance far from your system at which all the theory that we're used to use makes sense works is actually far. For example, for a five centimeter array or reflected array at one Terahertz, your far field, it's 16 meters far. What does it mean? It means that if you have a user closer than 16 meters and you're just using your traditional beamforming algorithm to point to that user, well, that's actually not going to work because all the beamforming theory that we have has been designed for the far field. Then if we're going to be operating in the near field, and again, I know it's contradictory, right? It's a little bit confusing. We're talking about small wavelength, but then large antenna structures, which means that the far field is far. So we're again in the near field. Fine. What can we do in the near field? Well, in the near field, we can do, for example, what groups usually you have done in optics, which is instead of talking about beamforming in that direction, you beamfocus on a point. And that's why in the last, I would say in the last few months, there are more and more papers talking about 6G in the near field, beamfocusing for 6G. And that's great. We can do that, but we can actually do something better. Let me show you in the next slide. What can we do? Well, welcome to the beautiful wall of wavefront engineering. What is wavefront engineering? Well, look, being in a communications groups, many times, you know, first of all, when we were at lower frequencies, we thought that signals just would go everywhere. Then, when we go to higher frequencies, we thought that we can send signals in different directions. And yes, we can. But there is another thing we can change of that signal in space, which is how that signal looks like as it propagates to you. So think of a signal coming to you. Usually, you just think of a beam coming to you. But what is that beam? Well, traditionally, we deal with Gaussian beams. So beams that have the typical shape of a beam, like a Gaussian. Fine. What's interesting is that, well, if you really have this large number of elements erased, like an ultramassive memory, if you really have that and you can really control your signal with more than half a wavelength precision, you can start doing different things with your signal. For example, once again, in your typical Gaussian wavefront, in the near field, you're focusing and in the far field, you're diverging. Fine. But you can do other things. For example, you can make a vessel beam. What's a vessel beam? A vessel beam, think of it as beamforming, but in the near field, which is interesting. Okay. So instead of doing beamforming with the traditional algorithms, we may want to do beamforming with vessel beams, for which we need some new theory. And that's great. What's also interesting is that vessel beams are self-healing. What does it mean self-healing? By the way, this is how a vessel beam looks when you look at it from the front. So this is the beam coming to you. A vessel beam has these rings, okay, and different orders of vessel beams have different numbers of rings. What's interesting is that as long as you don't fully blockage the entire beam, even if you block the center, even if you block half of this beam, you will be able to recover your signal. In other words, in this illustration, we show how a vessel beam is used. There is a blockage. The blockage is going to eliminate part of your beam. But again, because as long as not all the rings are focused, are blocked, you're going to be able to regenerate the beam. This is something which is well known at optical frequencies, and people dealing with optical systems have been using vessel beams for different applications. What we're saying is that now for the first time, at Terahertz frequencies where your antenna structures can be compact enough and still have a lot of elements, you can start doing these vessel beams. And for example, maybe you don't need intelligent reflectance surfaces if you already have a front, a wave front, that can recover itself from a blockage. Of course, there are other type of beams. For example, there are these 80 beams, which without doing anything, they will go straight, and then you can bend them. There is no device here. It's just a source on the left. Then you send the signal, and you can pre-program the signal to turn in space. Those are 80 beams. And again, for communications people, they are very new. For optical people, they actually, some of them have used them for sensing, for example. But we believe that this is something that it makes sense. This is something that finally should be adopted in practical wireless communication systems. And what's interesting is that, okay, when you have these wave fronts, and this is something that you will hear more and more, when you have these wave fronts, they also give you another thing to control, because when you add information to a signal, usually what you do, well, you can manipulate the amplitude, the phase, or the frequency, the polarization. Well, if you can manipulate your wave front, for example, you can throw in orbital angular momentum, which is something that you can manipulate to, which relates to, instead of having a beam that just comes to you, have a beam that turns on itself like in a spiral as it comes to you. And what's interesting is that different degrees of turning are orthogonal. So you can create orthogonal channels for your MIMO system. And instead of relying on the channel, you can rely on the emitter itself. So usually when we create orthogonal MIMO channels, we need the channels to be, you know, to go through that richness, that multi-path. What we're saying is that OAM, it's not a different dimension. No, OAM, it's just a way to create orthogonal channels directly from the transmitter. And again, these are opportunities. So when I see works on terahertz that relate to, you know, the channel is on that, I think that it's great. We can do another hybrid pre-colding beamforming paper. But I think that there are many opportunities that honestly, we don't know, we need to work on, and that are very exciting. Now, and I know I'm talking a lot, let's shift to communications and signal processing. Now that we understand that there is already many, there are already many things for you to play physically with the technology and the channel, but there are many things that we can still innovate. What happens with the physical layer? Well, look, the key thing at the physical layer is that, remember, we're going to the terahertz band with the promise that we can talk at terahertz per second, right? And if you talk at terahertz per second, synchronization becomes challenging synchronization in time, in frequency, and in phase, because you know what? Our oscillators might not be the best ever. Now you need to do channel estimation, equalization over large bandwidth. And again, there might be solutions that can do that better than with traditional systems, right? Especially you're trying to do very quick channel estimation and equalization, because look, if you tell me, yes, I can transmit at one terabit per second, but I need one minute to get my equalizer ready, or I need one minute to estimate the channel, then you know what? I'd rather go at 2.4 gigahertz and transmit at few megabits per second. It doesn't compensate. So you need efficient low latency solution. How about modulation and the modulation? Well, can we use the traditional modulation schemes? Of course you can. There is nothing that prevents you from using QAM, PSK, OFDM, and that's fine. That's probably what's going to happen, unfortunately, in 6G. But what's exciting is that you can do much more than that. You can do much more than that. And let me show you, for example, some of the modulations which are less traditional and that you can use for terahertz communication. For example, just think, right, we started with terahertz thinking of short range. So in the short range, what type of modulation we can utilize? Well, we propose a very, I would say straightforward modulation, but that works. We propose to utilize these very short pulses, pulses which are less than one picosecond long. In reality, they are hundreds of femtosecond long. And we make the system very simple. If you want to transmit a one, you transmit a pulse. If you want to transmit a zero, you just don't transmit. And why did we propose this? Well, because actually these pulses are very common in terahertz sensing applications. So if, eventually, you want to design a terahertz communications and sensing system, what's better than to use not only the same hardware, but also the same waveforms? And again, today, everyone will say, of course, it makes sense. Well, look at this. This was in 2014. And actually, the conference paper of this work was in 2011. So you can see that there was some resistance when trying to publish that paper because people were not necessarily ready. Again, these pulses are great, but mostly for the short distance when there is not too much absorption. And why? Because these pulses are very brought in frequency. So they would go over the absorption lines, which would mean that we would lose an excessive amount of power. For longer distances, what happens? Well, when you go to longer distances, what happens is that the bandwidth of your channel changes with distance. For example, here on the left, I'm showing you how the channel coefficients look between 900 and 1.1 terahertz. And what you see here is that if there were no absorption, this would be a flat line with a value of 1. But because there is absorption and there are absorption peaks before 1 and after 1.08, we see this kind of shape. This is a transmission window. What happens is that when you increase your transmission distance, your channel coefficient goes down, which means there is more attenuation. And you will say, yes, that's perfectly normal. What's new here? Well, now let's look at this figure on the right. Let's normalize these these curves, the blue line by the maximum of the blue line, the red line by the maximum of the red line, and so forth. And when you see that when you normalize these lines, what happens is when we went to 1 to 10 meters, we did not only lose power, but we are losing bandwidth too. To go from 1 to 10 meters, your absorption, your transmission window went from almost 100 gigahertz to just 50 gigahertz. I mean, note that I said just 50 gigahertz when 50 gigahertz is two orders of magnitude more than what we use in 5G. But this is something to take into account. Now, there have been different solutions to take advantage of this. Some of them are more complicated. Some of them are more straightforward. Recently, we have been working on this idea of hierarchical bandwidth modulations. So it's a single-carrier communication scheme in which you can allocate two simultaneous streams in which it's not that you just play with users having different SNRs, but you play with users having different symbol durations, so symbol times. So you can think of hierarchical bandwidth modulation as your hierarchical modulations or concatenated modulations that you can find in, for example, some digital video broadcasting systems, but in which we don't just add up the signal that we're sending according to the SNR, we add up the signal to the SNR and to the bandwidth, and we can actually multiplex in constellation different users. And again, we went from this being theory to recently actually implementing and demonstrating this with one of our testbeds. The results are in a paper that are still to appear, but this is how it looks like. A user who is close to the transmitter can recover a signal that looks like a 16-quam. A user who is far from the transmitter can recover a signal that looks like a QPSK. But still, they all have information, and what you need to do is smartly allocate some bits for the near user, some bits for the far user, taking into account the channel. Bottom line, molecular absorption is not always bad. Molecular absorption is, for example, what's allowing us to multiplex users in space. Now, I understand that many of you may be still thinking, but Joseph, can you give us the example of one application that needs one terabit per second? And it's a little bit challenging, right? I sit in satellite communications because you can't re-aggregate it flows. I sit in backhaul links, maybe not for the users. Maybe we don't need one terabit per second in a cell phone. But what can we do with a very large bandwidth? Well, one thing that we could do is to revisit some thoughts from SPET spectrum. And for example, we can talk about direct sequence of SPET spectrum, your CDMA, that thing that you use in the US in the 2.5 and the 3G and the 3G everywhere, which is the CDMA. What's the advantage of CDMA, or code division multiple access? Well, it's that it allows different users to use the same frequencies at the same time over the same space, as long as they use different orthogonal codes. The problem is that this direct sequence of SPET spectrum, it's not a very spectrally efficient modulation. And that's why we move away from CDMA. But actually, above 100 gigahertz, we have a lot of bandwidth. Maybe it's okay if we are not super spectral efficiency. And if we adopt direct sequence of spectrum, we can create like a CDMA above these frequencies, but also we can make solutions, we can coexist with narrow one users like a radar, or which can coexist with passive users like those that are listening to the signals coming from space. So we can do that. And we actually recently experimentally show some direct sequence of SPET spectrum system. There is another type of SPET spectrum that you can do, which is called the CHIRP SPET spectrum. CHIRP SPET spectrum. Some of you may be familiar with CHIRP. What's a CHIRP? CHIRP is a signal, it's a sinusoidal whose frequency changes through time. For example, in an up CHIRP, your signal goes from a low frequency to a high frequency. And in a down CHIRP, your signal goes from a high frequency to a low frequency. This modulation, it's quite popular, for example, in IoT applications. You may be familiar with LORA 1. LORA 1 utilizes a modified version of CHIRP SPET spectrum. And if you're a radar person, you're familiar with CHIRP SPET spectrum, because usually that's what you use, those continuously frequency modulated signals in your radars. What's interesting about this waveform? Well, this waveform, it's very robust against frequency selectivity. Why? Because the information, it's not on the amplitude of the signal, it's not even on which exact frequency the information is. Your information, it's on the trend of the signal. So as long as you can tell whether the signal is going up in frequency or is going down in frequency, you can already recover the information. Even if some of these frequencies are missing, that's great. So actually, could we use this type of waveform to operate even when you are partially over an absorption line? Well, that was pre-anxious. One of my students who is now in the market very soon proposed, and actually it makes sense, you can use this CHIRP SPET spectrum to communicate even when you are partially on top of an absorption line. So there are opportunities. There are many opportunities at the physical layer, and I will just mention some of them. What I show you here are results for when you are trying to make just one channel, but now try to couple these results with the results on ultramassive MIMO. Well, you need joint waveform and waveform design. That's the type of things that we should probably look. Of course, I understand that there are some people who are more on the networking side, and you will say, do I need to change my networks? Well, what happens here is the following, right? Usually, when we teach networking, we say the most scarce resource is your channel bandwidth, which is limited, blah, blah, blah. Well, when you go to Therahertz frequencies, your most scarce resource is not the channel bandwidth. Potentially, there is a lot of bandwidth for everyone, but what's difficult is to get everyone synchronized because everyone is using high gain directional antennas. And you cannot play the trick that at low frequencies you do even at millimeter-way frequencies, which is like, you know what? At least in the discovery phase, one of us can be omnidirectional. The other needs to be directional, but one of us can be omnidirectional. Well, at Therahertz frequencies, if one of us goes omnidirectional, we're not closing the link. We cannot find each other. So there are many problems, problems for which we already have some early solutions. And for example, we suggested that if you have an infrastructure network, maybe instead of going from the typical pushing model, which is like, you know, you as a transmitter have data, and you try to find the access point to pass it, maybe we need to modify, enhance, and go back to some sort of polling, meaning being the access point, the one that interrogates you. So instead of being a transmitter-driven MAC, have a receiver-initiated type of medium access control protocol, all implemented with, you know, under the assumption that you have beam steering antenna rays that can cover space with multiple sectors, this and that. We actually had some original ideas, then we actually doubled down on them and show that for reasonable parameters, you could get to medium access control or link layer throughputs that are in the excess of 100 gigabits per second. And you will say, why do you say that, Joseph? We thought we were doing terabits per second. Well, terabits per second is what the physical layer should be able to provide you. But remember, there is an overhead for you to get there. What we're saying is that we are trying to make our overhead as little as possible so we can offer you something that as a user will perceive as 100 to 100 or even 300 gigabits per second, which is, again, believe me, it's a lot, as at the link layer. Of course, there are, you need new ideas for neighbor discovery. And you can take advantage of how your antennas look like to do that. You can do better than just doing kind of a very sequential mode of discovering each other. But I'll just don't spend too much time here. Other things, if you're trying to communicate to someone who is far, you don't necessarily need to directly reach that person who is far. You actually may want to go through multiple hops and multi-hop relaying. It's a very well researched topic at lower frequencies. Do we need new studies at terahertz frequencies? Actually, we do because you know what? As we just explained, when you change your transmission distance, not only you change the power that you need, but you also change the bandwidth that you have. And there are some very interesting trade-offs between the distance and then the bandwidth that you can use, the bandwidth of your antenna, the latency that will be introduced because you need to find the next hop and all these things that are very exciting. So we very recently published some work in which we don't tell you exactly what's the definite solution for your system, but we provide a mathematical framework with which you can actually explore the trade-off and then maybe build how it's going to be your infrastructure network for this. Look, we keep saying that we have these solutions and we try them. There are two ways in which we have been trying the solutions. On the one hand, we implemented the channel models, the antenna models, and even the physical layer solutions into NS3. We package it all together in an open source downloadable app, which we call Terasim. And I'm telling you that again, we started this, but now there are many other groups who are contributing to Terasim and we're very happy with that. Please do that. But we also have an experimental testbed, the Terra Nova platform, which was funded with the support of both the National Science Foundation and the Air Force Research Lab. And in this case, what we are showing here, it's a system that can cover different frequencies from 120 to 140 gigahertz and then also around 1 to 105 Terahertz. These are all electronically driven devices. So no, they are not our Plasmonic devices because they take some more time to get there. But we have those devices and we have different types of antennas. And in terms of bandwidth, we can process all the way up to 32 gigahertz of bandwidth, which again, it's not the 50 gigahertz. It's not the 100 gigahertz, but it's something much more than what you have today in 5G. And I'm not showing you this slide just to make you jealous. I'm showing you this slide because look, we have already used this platform for many things. We've done some actual data transmissions of 1 Terahertz. We've even done multi-kilometer long links. We implemented this ultra-broadband spectrum. But what I want to say is that it's our mission to actually collaborate with you. And if you have some ideas of, oh, I would really like to track that new modulation. I would like to see what happens when the hardware is in the loop. Well, we're very happy to collaborate with you. We actually made the platform accessible. You can come yourself. If you have a group, you can send someone in your group to work with us. That's very, very fine. If instead you want to train your solutions with something that we already have, we actually have released a dataset with signals that contain data. So not just sinusoidal, but actually information carrying signals with well-defined frames. We have defined them. We have shared them publicly in a public repository that you can find from the lab. So our goal is to share because, again, we can do many things, but we can do much more by collaborating. And that's the goal. Now, just to wrap up, we are in an ITU talk. What happens with policy? Well, remember, for a new technology to go beyond the lab and to actually have an impact, we don't write papers just for the sake of writing papers. We write papers because we want to have an impact in society. What do we need to do? Well, first of all, we need the technology to be legal. We need to be able to legally use this technology. So we need regulations. The worldwide regulations are set by ITU. And for example, then in the US, the FCC adapts those worldwide regulations to the country, different agencies in different countries. That's great. For example, it was in 2019 that the FCC created a new type of experimental license that you can very quickly request if you want to test that technology between 95 gigahertz and 3 terahertz. In addition, they created some ISM bands and all over the place like, oh, the FCC makes 20 gigahertz of bandwidth available for testing in an unlicensed way. The problem is that it's not that they gave me 20 gigahertz of bandwidth. It's that they put 2 gigahertz here, 5 gigahertz there, 1 gigahertz there, which is great, but it doesn't make sense. I want large bandwidth, right? And there are other things that we need to study, which is how we coexist with those sensing users, including those that are on satellites. And that's why, again, we need regulations. The other thing that we need is that we need to agree on how we use the terahertz band, which means that we need a standardization. And there is this group in IEEE 802.15 in which in 2007 they approve the first terahertz of standard. I would say it's a very, you know, very straightforward standard. It's only for point-to-point links. Doesn't mean that that's the best thing you can do. No, that's a starting point. And mostly if you understand the context of that standard, that was mostly to get industry into the loop. Because once there is a standard, it means that if someone makes a product that it's compliant to the standard, it should be able to work with other devices. So this was mostly to attract industry. Do not think of this standard as the best thing that you can do with terahertz. It was a starting way. And you can join the group. This is an open group, and you can participate in that. With this, I know I went well over time, and I apologize for that. But I just want, as I said at the beginning, right? I started on a path with Professor Achilles and I tried my best to grow. And of course, this is not just myself. It's a bunch of very talented people with whom we're doing the work. With this, I think I'm ready for questions. Thank you, Giuseppe. It's fantastic. I didn't want to interrupt you, but I think it's just a couple minutes above your time. So we started a little late. Fantastic, really. Thank you. And there is a question by Feroz Sageschi. He says, thanks Giuseppe for this very iterating presentation. I don't know what he means by iterating. I think maybe interesting. Interesting. Yeah. Yeah. Right. What's your view on the potential application of terrorist communication to connect UAVs? It's a really very general question. I mean, it will take you. Don't take more time because what are the challenges that need to be solved? So I would say that you can write a paper on this. Yes. Well, what I would say is the following. So again, I've seen some papers on UAVs and terrorist. The main thing is that you're dealing with very high directional antennas at the transmitter and at the receiver. So if you can precisely point and have both transmitter and receiver point it, that's great. That can be a way to go. The other thing is that you could rethink how you want to be connected to UAVs. And the way to rethink that is like, look, when you are at terrorist frequencies, if you have the channel, the channel can actually be very good. And it can allow you, it allows you to dump a lot of data. But a lot of times you will not have the channel. So maybe you should start thinking of creating networks that by default are not connected. Only sporadically. Sporadically means every few seconds you get connected with the promise that when you get connected, you get this huge pipe to throw your data in. But again, mobile terrorist communications in general, I guess that it will all come down to how you keep awareness of where everything is. And that's why maybe you don't want to just do communications. You need communications and sensing when that sensing can be tracking. That's a short answer to what could be, you know, probably a long paper with tons of challenges. Yeah, there is no question, but I have some questions so that we can make it more interesting. You did not talk about the synchronization. I mean, you mentioned it many times, but somehow there was no time to go into details and also the equalization. Those are very important especially for terrorists. What is your opinion about those two problems? We did some work on those two back in 1415, as you remember. So what is your status now on that? So since your papers with Chong, there have been not many works. Actually, I would say there have not been any work really dealing with physical layer synchronization. And not because it's not a problem, but actually because it's a problem that until people get starting playing with the devices, they will not see that it's an actual problem. I would say the following. Why do we need synchronization? Remember, you're dealing with clocks that are not stable. So there are all these, there is this phase noise and jitters. You're dealing with weak signals and you're dealing with, again, fluctuating phases. So because of that, you need to get somewhere better. I can tell you that what we have been doing these days is actually characterize the phase noise of the devices that we have right now in the lab. Of course, the devices we have costed, some of the components are $80, $90,000. So they are very good even in terms of phase. To get synchronization, we are dealing with different type of synchronization preambles. We can reuse some of the lower frequency ones, but we need to do better synchronization preambles that take into account what really happens at terrorist frequencies. Perhaps even bigger problem or, I don't know, it's to do the channel estimation and equalization. Because right now I can tell you the majority or actually all the experimental works that have been shown out there use the very traditional channel equalization like the MMSE or ML equalizers. And what no one says is that, well, to estimate the channel for a few milliseconds, you probably are consuming 20% of your frame. So you think you're sending data, you're just sending pilots. So you need a lot of pilots to be able to estimate the channel. One thing that could be done, and I know that some groups are working on that. I'm not sure if those works are out. Can we take advantage of machine learning as a way to quickly understand the channel? This is something, however, that then opens another problem, which is like, are there processors for machine learning algorithms that are fast enough? Because I see many people who talk about machine learning at terrorist frequencies, and then you say, okay, how would you, what type of processor would implement this, your machine learning network, right? What type of processor would implement your neural network and can digest the data as fast as it comes? Because again, things are coming fast. And that's a little bit of a challenge. That's why these days there are, I'm aware of a few startup companies who are actually working on accelerators for machine learning algorithms, who could, who can then be utilized for real time channel estimation of this large bandwidth. But I would say, you're right. There have not been many works since your works with Chong, I believe in 2014 and 2015. And I think that as people get more experimental and they see that this is not just a paper to write top, it's not just a topic to write papers, but it's a real problem. We're going to see more. I have to mention also the credit to Wolfgang Gastaker from Erlangen. You mentioned, yes, right, right. Wolfgang was a key player in that, especially for the equalization, we call them mystery equalization. So why is it also important is from the receiver perspective, right? You have to somehow detect the incoming signals, they are coming extremely fast. And now you mentioned about machine learning, I'm afraid that the entire world will run hundreds of papers on machine learning for terrorists. Maybe they're already there. Remember, we had also some papers about machine learning for terrorists, which Ryan, Ian, you. So one more, there are more questions coming now while we're talking. There is by Sabit Ekin, he says, great talk. What are your thoughts for FSO? I'll say it very quickly if you want, right? So look, yeah. So difference between terrorist and free space optics, what I would say is the following. Yes, at optical frequencies, you have even larger bandwidth than a terrorist frequencies, but those comes at the cause of even more propagation channels, challenges. So right now at terrorist frequencies, you already get which are larger than what your hardware can process. And when people say, oh, but the terrorist channel is not good, I say, excuse me, look at the optical channel, right? It's even more challenging. So what I would say is that there are applications like, you know, perhaps even above the atmosphere, perhaps when there is good weather, you can use optics. But remember, with optics, you're dealing with pointing lasers, which is even more challenging than pointing antennas. Yes, you will have more bandwidth. The reality is that the bandwidth will be difficult to utilize, not because of the channel, but because of the technology. So I would say we will eventually get to optics. I think that right now, the sweet spot for what we need in terms of data rates and what we can do with the technology, or not going to get the technology, but what we can do realistically for the applications we have, it's a terrorist ban. So there is another one by thanks. So Masoud Ashgari or Ashgari, thanks for the presentation. Do you agree that we need much better energy consumption models for terrorist transmitters and especially the receivers? Current papers just use your assumption of on 0.1 energy use of the receiver compared to the transmitter. Now I remember our paper back in 2011, right? Yes. I cannot cover everything. Please. So this is a very good question. For those who are not familiar, so yes, we didn't talk about energy consumption, right? And usually if we mention anything, is how much power the transmitter consumes? Because at the transmitter, you need this power amplifier that can get your signal as strong as it can outside. At the receiver, what's the power consumption? Well, the power consumption is certainly lower because you don't need that power amplifier. If anything, you will need a low noise amplifier, but usually those consume, yes, one order of magnitude less of power than the transmitter. I would say you can go with what we said back in 2011 and 2012, assume that your receiver is one order of magnitude less. And what I would say is that if you're thinking of that model wasn't the focus of nano communication networks. If you're thinking more on the context of large scale or macro scale telehertz communication networks, there is another thing that consumes a lot of power and it's your digital back end. These analog to digital converters and digital to analog converters, they consume a lot and they are both at the transmitter and at the receiver. So I would say still your transmitter consumes much more, but actually they consume quite. So we have been so driven by just talking about longer distances and higher data rates that we have put the power problem a little bit into the drawer. Maybe it's time to open the drawer and look again into a revised energy models for Terahertz communications and sensing networks, not only nano also macro. I see another question. Sometimes when I see these questions like, I don't know what to say because he will be shot now. The question is what kind of pseudo random codes do you plan to employ to produce preamble during the synchronization? So good luck to answer it. I mean, okay. So to improve synchronization, what do you need? You need a code that has very high auto correlation and very low cross correlation. There are codes like that. And actually you can look into some of our papers. We actually say what type of code do we use for synchronization only and there's a maximum merit sequences. I can tell you that we're doing many things just with 11 bits. The other type of thing is what type of codes would you like to... So those are codes for synchronization. Another question would be what type of codes do you want to use for error control? But that's another problem for another day. Yeah, so I think I would like to discuss some more interesting things nobody asked those questions. Maybe offline but then there are all these people and maybe these things will give them ideas. So I like or I love this idea about using the chirp spectrum for the ultra, you know, ultra, you call it ultra. Because remember the Laura was using it for the distance combating, right? Yes. I did not get the novelties but we can talk offline. So I love that. And then also I love the, you know, your approach to protocols. But I think it's time to do some machine learning empowered protocols. And then also like, you know, do we really need the old fashion of these protocols? You know, there's this Reinhardt doesn't like that but there's this new IP ideas are floating around. And, you know, maybe we really need to rethink about this question for the terror expand. Like, you know, can we eliminate MAC, for example, routing or I don't know, so because we have all these new technologies, right? So, but you know, I'm expecting more very pioneering work from you and your group. So thanks a lot again, Josep. And I will ask Alessia to jump in. And many people loved your presentation. And so I ask Alessia to take over. And again, I thank you a lot. Thank you, Professor. Thank you very much. Professor Achilles for moderating this session. I can see there are other questions, but it's really late now. We need to move to our wisdom corner. Thank you so much for your very informative presentation, Professor Jornette. So welcome to the wisdom corner, a live life lessons, which is based upon the idea to give a unique special angle to these webinar series, as we said, a personal touch. So successful researchers like you today will guide students and young scholars in the field of current IZT research, but also we'll try to share some impactful life lessons. So I would start with with my first question. So which is your your hard earned life lessons or failure that you would like to share with us today that might perhaps help somebody attending this webinar? Absolutely. So look, I think that this was already a little bit mentioned at the beginning with Professor Achilles introduced me, but what I would say is, when in 2009, right, we suddenly start talking about Thetahertz, right? And Professor Achilles is convinced that this is going to work. But I remember that when we wrote the antenna paper, we got the conference accepted. And we send it for a journal publication. And the reviewers came back and they were like, really bad. Like, you know, what are you talking about? Graphene is a very new material. No one really knows who are you to even think that you can use it. Like, almost personal attacks. I said, what have we done to these reviewers? So it was hard, you know, you were we were working on a topic that it was very new. And as a first year PhD student, well, you you can either be lost or have a very supportive advisor. And that's what I had with Professor Achilles. He kept telling me, Josep, don't even read those reviews. I mean, read them so you can learn something out of them. But put them in the drawer, we'll get there. We'll get there. And that's what happened. But what I want to say is that through a PhD, and even later on, but through a PhD, I mean, you know, you are going to go through these ups and downs. What's important is that, remember, your advisor is there for you, and your advisor eventually will become your mentor. So it's there. Do that. Don't get discouraged by reviewers, even if they say something that attacks you, that's very bad. Unfortunately, we have people like that in our community. But don't lose faith. If what you're doing, it's right. If you have the support of your team, of your advisor, just go on with it. And again, 2009, people thought that this was science fiction. We can see that it's not. Thank you. Thanks a lot. It is not. It is reality. Which strengths you think and capabilities students, young scholars and researchers should be most focused on developing? And how do you think they should plan on accomplishing this? That's a very good question, too. So what I would say is the following. We are in today's society, there are, you know, it's very easy to publish papers, right? So there are thousands of conferences, journals, this and that. So as a student, right, you will say, oh, I want to know everything, or you try to catch up everything. Well, you know what, you cannot. And in fact, you should not. As a PhD student, what you need to develop is depth. You cannot just hop from different things. You need to pick one topic. And of course, you agree that with your advisor, but you need to pick something. And in that something go in depth. Don't don't. You need, okay, I said go in depth because you need to build skills. You need to don't be afraid of going to fundamentals. Like, for example, to do our works on the antenna or on the channel. We took quite the tour in physics, right? I remember my first semester at Georgia Tech. In addition to courses on protocols, I was taking courses on quantum mechanics and molecular nanoelectronics. Why? Well, I remember walking back from one of Professor Achilles classes back to the lab. I said, I don't know which other class to take. There is these are the options. And he said, oh, you should totally go to this quantum mechanics and molecular nanoelectronics class. So I went, I was the only one from comes there. But that's what opened the door for me then to be able to start doing this antenna work, this channel work, which requires physical knowledge. So what I would say is what you need to be able, you need to be able to know how to focus. You need to be able to understand what's going on. You need to know what's going on in your community. But you need to have your home area. You need to have your depth knowledge. That's where you're going to, that's how you're going to succeed. Thank you. And to go deeper, which fields and which topics would you recommend students to study nowadays? Well, so first, I would say, because of my experience, what I always say is like, look, the only thing that the only thing that you know, it's true are the laws of physics. So in my case, even if I'm an engineer, many times I will end up taking the tools and try to learn more about the fundamentals of physics and apply them to electromagnetics, apply them to electronics and apply them to communications. I would say that recently, there is an initial trend and I think it's going to grow, which is these people who are trained at the intersection of electromagnetics, communications, and signal processing, because it's not so longer just about doing waveforms, it's also about doing wave fronts. So I would say that is the intersection of the fundamentals, some physics, some electromagnetics, some signal processing, that's great. And then what do you use it for? Well, you can use it for your favorite type of network. One thing that we're experiencing as academics is that we don't have as many students, as many stop students that we used to have in the past going to communications, because you know, the word nano or bio, it's more catchy. And what I always say, you know what, you can do that as an engineer. And for example, or the internet of things, internet of nano bio things, the nano bio communications and sensing are real topics. And to people who say, who in 2009 said, Terahertz is crazy. And today, they may feel a little bit like, oh, these nano machines in the body or in the brain are crazy. I don't think that they are more crazy now that Terahertz was 10 years ago. And if you want to make a contribution that will last, that's the type of crazy topics you want to do. So you need the tools, study, go in depth, and then use them for something that not everyone else is already doing, even if it sounds crazy. Okay, even if it sounds crazy, as long as you don't grow against the laws of physics, you're fine. It's going to happen. Thank you. Tell us one of the most tangible contributions that you have made in your career. I'm sure there are many that had a direct impact on your life or your professional or even personal life that you're most proud of. So it was difficult to pick one. I would say, look, one of the key works that is now being very heavily cited and that it's nice, it's the one hand, it's the Terahertz channel model. This is a work that we started in 2009. And I would like to say with this Terahertz channel model, first, we have to convince ourselves, right? I mean, will the Terahertz system work? But it's a very mathematical, very, very physics based model that it's at the basis of many of the things that we're seeing today. It's at the basis of how you can use Terahertz for short range. It's at the basis of all these absorption windows in the long distance. It's at the basis of this different type of signal processing solutions that people are doing. And we're happy for that, right? It took us some time and some sweat, and it got out. And the other one, if you allow me, it's the nano antenna work, right? We started working on Terahertz because we wanted to create nano communication networks. Usually, the papers before our paper were about, oh, you cannot use electromagnetics because all of these problems. And we said, no, actually, you can make an antenna, which is a small and which we can, which can work at these low frequencies compared to what the other systems are. So Terahertz compared to what they were suggesting, it's low. And I think that that was the enabler for many things, many things that we're doing today. Thank you. If a professor would like to intervene, feel free. If not, if you have any question, you would like to ask Professor Jornet, please feel free. If not, I would go to my last question. Okay, so before we close this webinar, is there a motto, an aphorism, or a book, a movie, a piece of art or music that you believe describes you best or probably your professional path that you wish to share with us? Okay, it's actually something that I already said today, which is like, and I realized by the end of the week that that's something that I repeat quite often, which is like, as you know, the only laws that we cannot change are the laws of physics. For everything else, it's just a matter of time, right? That's that applies to technology, that applies to how we behave, right? We are in a society, we know how to negotiate, how to interact. The only thing we will not change are the laws of physics. And again, if then you allow me, then some of my students can say, well, we can always look into another set of physics, fine, but again, the motto stays, you know, the only thing you cannot change are the laws of physics. Everything else, we're going to get it. Thank you. Thanks a lot. Professor Akildi, if you want to jump in now. Again, I thank you, Alessia, for taking care of this session. And I wholeheartedly thank to Josep for this outstanding one and a half hours. I really enjoyed it. I missed our times. I wish we could go back again and really do those things. But what can we do? That's so, so it's life. And now you have your own kids, meaning not only your own kids, but also your PhD students. It's lovely to see that you're producing excellent results. And I expect more from you. Again, thanks for participating to our program. And I would like to announce also that we lined up outstanding speakers as like the first phase in the fall term. It will start in September. And we will announce the program in due time and hope to see you in fall. I wish you a super summer. Stay safe and healthy. Be careful with COVID. A lot of my friends who are vaccinated are still getting the COVID. I just found out from my colleague in Iceland. We had three shots and we still got to COVID. So just stay safe and healthy, please. And continue to work. And I hope Josep's talk as well as the other talks in our program gave you some ideas for front, for front problems. And please tell to all the younger people do not work on old horses, meaning a lot of researchers are wasting their time on working on old problems or same problems again and again, just to write papers and get their promotions and tenure. But it's really a waste of time and also cheating yourselves. Work on really hot topics, try to advance technology. You see Josep's really like the last 15 years, his entire career, how he developed things. And now he's now pushing for the test beds and physical implementations. And so that's a very good example, honestly. So thank you again, Josep. Thanks a lot, Alessia. Thanks a lot, Reinhardt and Erica and all the other ITU colleagues. I wish you a super summer. So thank you again. Thank you so much. So thank you, special thanks from the ITU as well to the speaker, Professor Jornette. Thank you. Thank you so much. And you mentioned already, I would like to thank specifically as this is the last one of the first series, the director of the telecommunication standardization bureau, Dr. Chesa Blee and my boss, Dr. Reinhardt Scholl, the deputy director for their continued support and trust. And of course, Erica, without her, all these would not be possible. Thank you, Erica, for your outstanding support from the ITU journal team and my colleagues for IT and logistic support, Gifty, Gant, Carlos, Elia, Danai, everybody. Thank you so much. So we look forward to seeing you all in September. Bye-bye. Bye-bye. Ciao, ciao. Thank you. Bye-bye.