Modeling Reflective Intelligent Surfaces (RIS) Systems with MATLAB

Hello, everyone. My name is Houman Zarrinkoub. I'm the Principal Product Manager of the wireless communication products here at MathWorks. And today, I am joined by my colleague and friend, Rameez Ahmed.

Thanks, Houman. And my name is Rameez Ahmed. I'm an application engineer here with MathWorks, and I deal with all of our wireless and wireless standards based products. So anything to do with 5G, LTE, wireless, LAN, Bluetooth, and satellite communications, those are all my areas of interest.

And we would like to welcome you to this MathWorks webinar entitled "Modeling Reflective Intelligent Surfaces with MATLAB." Let's go over the agenda of this webinar. Following some introduction, we're going to define the reflective intelligence surfaces and their place in wireless communication, talk about the benefits and use cases, design challenges. And then we're going to go over the RIS modeling in MATLAB, and that includes discussion of the overall assessment, including channel models. And finally, we're going to summarize this webinar.

What is the goal of wireless communication? It's ubiquitous connectivity. You want to be connected no matter where you are in the world, using different types of networks, Bluetooth and low-range networks, local area networks, Wi-Fi, wide area networks, the 4G LTE, 5G, 6G that's coming up, and non-terrestrial or global area networks featuring satellite communication.

Now, when you talk about wireless connectivity, there are two metrics that are usually discuss. Wireless engineer try to optimize wireless capacity and wireless coverage. The coverage relates to the reach of a signal, the range of the received signal. And the further you can reach, the higher your coverage.

Capacity, on the other hand, relates to the number of users and each user, how much data rate we can allocate and how much bandwidth we can allocate them. Usually, it's a compromise between maximizing capacity or maximizing coverage.

Let's say you want to increase the coverage of your wireless communication system, especially 6G is coming, and the notion of ubiquitous connectivity is a goal of 6G. To increase the coverage, what are your approaches? You can add base stations. Make it smaller cells. But if you do that, then because the cells are smaller, you may have intercellular interference.

You have to incur the cost of putting new base stations, power, and so on. You can add relays or repeaters. These are active systems that take the signal, amplify it, and transmit it out. So that's power intensive, and the placement provides challenges.

Or alternatively, you can add controllable reflectors, not active elements, but controllable reflectors that are passive, low-power. And RIS, or reflective intelligence surfaces, is one of these types of systems.

Now, we are right now in the 5G era of wireless communication. Next generation or 6G is coming. And chances are the following enabling technologies are going to be heavily used in the 6G. 6G is not standardized yet. The requirements is not clarified yet as of this year, 2024. But when you read the literature, and they notice that these technologies are deemed important for 6G.

One of them is AI/ML, the use of alternative frequencies, millimeter wave, sub terahertz, mid-range frequencies, and reconfigurable intelligent surfaces is discussed among them, non-terrestrial network and cell-free massive MIMO. So definitely, RIS is one of the technologies that relates to 6G.

Now, how can we define reconfigurable intelligence surfaces? So let's imagine you have a transmitter you can see here, and you have a receiver. And the line of sight is not available. There is a blockage. You don't put a repeater, but you put an array of controllable, passive, low-cost reflecting elements. So the surfaces of buildings and other areas can be made of antennas or metasurfaces that are passively reflecting your signal.

So each element of that array can be reconfigured and apply a custom phase shift right to the incoming signal. And if the choice of phase shift is judicious, and you control it properly, they add up constructively, and you have more power at the receiver. Hence you increase the coverage. So there are two ways to construct the RIS, either using antennas or metamaterial.

And essentially, when you look at the gist of why RIS is so useful in terms of the next generation of wireless systems, because the RIS systems allow you to control the reflection angles. Look at the traditional way of reflection. If you have a traditional material with no controllability of its electromagnetic properties, you have a incident wave, could be like a magnetic wave, could be other types of waves, at an angle AI. By Snell's law, the reflected angle will be equal to the incident angle. So that's a deterministic reflection angle based on laws of physics.

Now, if you have a metasurface, and this metasurface is controllable, although the incident angle is still AI, but you can configure the reflection angle. So if you are trying to reach somebody who is behind the blockage, by changing the reflection angle, you can achieve that easily if you can control the electromagnetic properties of the surface where the reflection is happening.

So one of the things that is important for next generation of wireless systems is we are going toward the MIMO and massive MIMO systems. And when you have massive MIMO system, the beamforming or the directed wireless communication has a narrow beam width. So in massive MIMO, if you have line of sight between the base station and the receiver, and massive MIMO achieved through active arrays, digital antenna arrays on both transmitter and receiver. Now, beamforming improves the SNR and capacity and the coverage.

What if you don't have line of sight? There's a blockage. You don't have line of sight. What happens? If you put that array of risks or reflective intelligence surfaces, you can even achieve good coverage and good capacity even with non-line of sight scenarios. So you have a passive set of arrays, non-active, doesn't use too much power, in the surrounding areas, especially in the urban suburban areas with lots of buildings, which can be modified to have those metasurfaces. You have those available, and the beamforming can improve the channel characteristics and give you a higher coverage in a non-line-of-sight case.

Let's look at the example. Here you see I am using a MathWorks RF propagation indoor propagation capabilities to essentially clarify the notion of what RIS can do. You see that a radio transmitter, could be a Wi-Fi access point or a cellular hub, that is in red, and that's a transmitter. Behind that, there is a wall.

Most probably, this wall is a blockage, have some metals in it and can block the signal. On the desk in blue, you see this small cell phone or other receiver. Because of the blockage on that wall, there is no line of sight. But if you have those reflective intelligence surface on the other side, on the wall, on the other side of the room, you see that the incident signal comes in. And by controlling the properties of the RIS surface, we can have any arbitrary reflected angle so that blue receiver, thanks to RIS, can actually have a very nice receive power. That's the story that RIS is provided.

The same example can be shown in outdoor propagation. Now, is this only applicable to communication? No, RIS can be applied to both communication and sensing. And that's great because in 6G, one of the applications of 6G is deemed-- we can talk about that soon-- the integrated sensing and communication. So if RIS can be applied to both the communication and the sensing module, that's great. So we can get the benefits from this in both cases.

And that's what I'm talking about. The integrated sensing and communication is one of the 6G application areas, and it's described as this. In an ISAC framework, both sensing and radar and communication functionalities are integrated with the same hardware and same transmit waveform. They're using different times. So ISAC framework have characterized the parameters, the location of a target. So when you're using radar or some kind of a positioning localization, you find the location of target, and then the RIS takes that location information and uses beamforming with the controllable angle of arrival and departure to beam forward toward where you want to go.

Now, I'm going to hand off at this point to my friend who's going to show you how we can put in MATLAB a nice simulation together that looks at the impact of the reflective intrusion surfaces on the overall performance of a wireless communication link. In this case, we may have a channel, we apply the RIS, and we see with RIS enabled and RIS disabled what kind of a increase received power and therefore, what kind of received performance we get.

Thanks Houman for that great introduction to RIS. Now we are going to start looking at some examples of how to implement RIS in MATLAB. And to begin with, let's start by looking at a very simple example. And so to make things very, very easy, we're going to use a deterministic channel model. And for the RIS itself, we're going to consider a 10 by 20 omnidirectional elements, each with half wavelength spacing, a carrier frequency of gigahertz. And we're just going to assume a SISO system, so which means that we're going to have single input, single output.

For the signaling methodology itself, we're just going to use very simple BPSK. So before I get there, let me switch into MATLAB and show you how to get to this example if you were to do this today at home. The simplest way to do that would be to use the phased array Toolbox. So you go to the doc phase to get to the Phased Array documentation page, and all of our examples are listed under this examples tab right here.

Now, if you go deep in and you can see all of these examples categorized right here. And for application areas, we're going to use a particular one for wireless communications. And if you scroll down, the example that we are interested in is the introduction to reconfigurable intelligent surfaces. I'm just going to copy this command right here and then go back into MATLAB and open up this example.

Now, we're going to walk through this example real quick, and then we'll run it and see the results. Now, what's happening here, like I said, we're going to set up a carrier frequency of 28 gigahertz, and we're going to have a 10 by 20 element resurface that has half wavelength spacing. Now, here's the RIS element itself being set up.

And let's talk about the scenario here. The scenario that we have here is a transmitter that wants to talk to a UE. And it has both line of sight and a path through the RIS. Now, the first scene that we want to set up is calculating a reference signal. So that's what's happening right here where we go ahead and calculate the channel between the or AP, whatever you call it, to the UE. And you can see that it's gotten a SNR of 18.5 DB.

Now, let's go ahead and add the RIS into the picture, and we go ahead and calculate the SNR again. And this time we're not really doing anything with the phase itself. We're just adding the RIS, and this is going to act as a reflector. Now you can see that it has a very minimal effect on the SNR. It's going to add the SNR. It's barely higher than the line of sight path.

So let's do something more optimal. Now, in addition to just adding the RIS, we're going to do some channel estimation and then set up some steering vectors, so that the RIS has an optimal phase shift added to the signals. So that's what's happening here. So we get the steering vector, and then with that channel information, we're going to set up the optimal phase controls, and then we're going to rerun that simulation that we just had and calculate the SNR.

And now you can see that the SNR for the line of sight plus the RIS situation is about 28.4 DB. So that is a significant improvement in SNR compared to the line of sight or the just adding a RIS as a reflector. So this kind of emphasizes the importance of having algorithms that can actively control the optimal phase.

Now, to signify the fact that RIS is actually doing a lot of important things, let's go ahead and rerun the simulation. But now we do it without that line of sight paths. Until now we had the line of sight path. But let's repeat that experiment but have no line of sight path.

So here's that calculation again. We go back and calculate the SNR in this situation. And you can see that it has a significant improvement over the line of sight path. So this basically says that even if you have just a RIS and no line of sight path, which is really not uncommon in today's urban scenarios. You could be walking around and lose line of sight with your cell stations or your node Bs.

And we also want to see what kind of impact the size of the RIS elements has on the data rate itself or any metric of performance. In this case, we pick data rate, but you can see that this is the line of sight path. And so that's remaining constant irrespective of the number of elements in the RIS.

But you can see that there is an optimal number of elements up above which the RIS starts to perform better. The line of sight plus RIS is really good, but you will start to notice that the RIS line itself for large array starting to converge with the line of sight plus RIS, which sort of explains, again, the fact that having just a RIS element is probably as good as having the line of sight plus RIS, which means that this works great in urban environments where you cannot have line of sight. You can also see the positioning of the RIS along where we were placing it. And you can see the SNR performance as a function of those RIS positions.

So this is a great introductory example for setting up a RIS element and performing some of those SNR analysis. Now, given that we were working on this in a couple of assumptions were very basic. The first assumption that we made was that this was a deterministic channel model.

Now, the example that we saw previously is a very simplistic example. It has a deterministic channel model, just uses BPSK and doesn't really represent a real world scenario. So let's take a more complicated example that uses a 6G like waveform, because as we heard from Houman, 6G is going to implement a lot of RIS elements. We're going to use an iterative algorithm in this example to control the phase of each RIS elements, and we're going to use a two concatenated channel models. These are defined by the 5G and R standards itself. So we're to use those channel models in this example.

Now, before we do that, I do want to point out that we need to be able to access the 6G Exploration Library, and you can get that Exploration Library by going into the Get Add-Ons tab button right here. If you click on Get Add-Ons, and you can search for 6G Exploration Library right here and install it. This is the Exploration Library that you're looking for.

Once you've installed it, we can come back and look at the documentation page for it. And here is the Exploration Library. We're going to look under the examples. And the particular example that we are interested in today is this model, reconfigurable intelligence services with the CDL channels. So let's go ahead and copy this command, and we'll open up that example.

Now, this example is very similar to the previous one, except now instead of having those deterministic channel models, we are going to have the CDL channel model in between the transmitter and the RIS elements and also between the RIS and the UEs. So let's go ahead and set up the scenario. You can see that we are setting up the scenario with some 6G like waveforms. So we define the carrier and the PDSCH properties for that waveform.

Next, we want to define the RIS elements. So we go ahead and define all of that right here, including the spacing for it. Now next, we want to define the channel model for that. So that's what's happening here. And the next thing we do is calculating the RIS link path loss. Now, this link path loss equations actually come from the ETSI report on reconfigurable intelligence surfaces. And it's part of the reference right here in this example If you wanted to take a look at it. It's at the bottom of the example.

All right, so we go ahead and calculate that link path loss. And the next step is we want to go ahead and calculate the RIS coefficients. Now, RIS coefficients is a local function right here that you can very well walk through. But you can see that if RIS is enabled, we go through a number of iterations and calculate the steering vectors and the channels and do the overall channel response for each iteration. And we use this metric to check for convergence as well.

Now, so that's the achievable SNR plotted as a function of the iteration numbers. Finally, we go ahead and generate the transmit waveform. And once we've done that, we go ahead and pass it through the channel that includes the RIS and go ahead and calculate the receiver side. We receive the signal and then go ahead and decode the PDSCH so that we can calculate the SNR and the EVM of the signal so we can see that it's got about 19% EVM.

Now, I do want to show you what happens in this case. So in this case, we did enable RIS, and that's how you got a pretty decent signal in there. What happens if you don't? Because remember, this is not a line of sight, so there's no line of sight path. Everything is going through the RIS. So if you don't enable that, what happens in this case? So let's go ahead and rerun that signal. And you can see that, there's really nothing received at all. So RIS, along with its optimal phase element calculations, leads to a pretty good SNR and also EVM in this case to establish that 6G link.

All right, so those are the two examples that I really wanted to walk through today. So one was a deterministic channel model. Another one was a stochastic channel model. One uses a very simple BPSK . The other one was a 6G like waveform that is being used.

Now, what other examples of RIS are out there? We also have a full electromagnetic analysis for that intelligent reflecting surfaces. This goes into the antenna toolbox, uses individual elements, and does a full electromagnetic analysis if you wanted to. These are the two examples that we looked at today. And in addition to that, we also have a radar sensing example. It's an indoor scenario trying to detect a human being walking around. And that's an example.

This is very new. It was introduced in 2024b. So it went out literally a couple of days ago. That's all the MATLAB examples that I have to show for you today. And now it's back to Houman.

Thank you, Rameez. As Rameez mentioned, in the release 2024a and now 2024b of MATLAB, we have introduced 6G Exploration Library. It's the collection of MATLAB functions and apps to let you explore, model, simulate, and test candidates 6G waveforms and technologies. It's an extension of 5G Toolbox. So if you have 5G Toolbox, you can download that for free.

And everything in that library of MATLAB function is full MATLAB open source code. Every function can be opened up. You can look at it, you can change it, you can modify it, and so on. The enabling technology of 6G, which is being discussed, it's not standardized yet, but being discussed, can be explored. And these include, as I mentioned, new frequencies and waveforms, integrated sensing and communication ISAC, and reflective intelligence surfaces for wireless and non-terrestrial network use of SatCom, or satellite communication for wireless communication.

There are multiple ways you can download that. First way, as Rameez showed, can go to add on and put 6G and find the 6G Exploration Library for 5G Toolbox. Or alternatively, go to your Google and your favorite search engine and put 6G Exploration Library, and you will find that library under MathWorks.

You can also go, if you have 5G Toolbox-- or if you don't have it, you have access to documentation for free-- go to mathworks.com/help/5G. Find the 6G Exploration Library section of the 5G documentation. And finally, you can go directly you can go to mathworks.com/matlabcentral/fileexchange and put the word 6G and find the actual place in our file exchange where you can download the 6G Exploration Library and install it.

So how do we learn more about all of this? Everything is within our wireless communication product pages. Especially 5G has a lot of 6G and RIS stuff in it, and please consult our main wireless communication solution page at mathworks.com/solutions/wireless-communications.html. To get more information about everything wireless at MathWorks.

To summarize, applying reflective intelligent surfaces to mobile communication can improve coverage, especially capacity, by enhancing the radio environment and controlling the propagation of radio waves through those controllable reflective surfaces. And there are new RIS capabilities in MATLAB that enables you to analyze response of RIS using interactive simulation, model RIS with deterministic and stochastic channels, and simulate cellular propagation environment in the presence of blockages with and without RIS. Thank you very much.