Wireless Connectivity and Scenario Modeling

Well, thanks for providing this opportunity to address the audience here. So I'll just briefly touch upon some of the future trends in the wireless space that we are foreseeing actually as we go. Saankhya, being one of the companies which have been in this particular space for quite some time. I wish to share some of the things that we foresee as some of the possible things that are happening in the wireless space, and also specifically what exactly the 5G also is doing in this as we move along on the different releases.

So just to--

OK. So yeah, this one just gives you an overall view of how exactly the migration of the wireless communication has happened over the years. So as you see, we started off with mainly analog radio signals, mainly working on the voice only section, and then migrating slowly towards the 2G, 2.5G, were actually the data part also came in one of the primary things, as far as the wireless technology is concerned.

With the GPRS/EDGE coming in, then moving towards the 3G part, which was strange for quite some time on 3G, where we had support for video calls as well as mobile TVs and video on demand, things came in. And much higher data rates came in with 3.5G, which the rates going up to 42 megabits per second. And 4G, as we know, is all prevalent right now, which is actually having an IP based networking thing as well. Right.

So and then the 5G space, which is actually the current resource what is actually there, as far as 5G is concerned, we are looking at much higher bandwidth and higher data throughputs that are actually coming in into the system. And also, there are things like low latency communication and UHF frequency communication also converging as part of the 5G systems.

So what exactly are we looking at as trends in the technology space here? So basically, we are-- I would say there are mainly three things. One is virtualization of the air interface.

In the virtualization of the air interface, we are looking at the-- we are seeing things like auto Open RAN communities and Open VRAN communities, actually defining things wherein actually you could basically clarify certain elements of the 5G ecosystem. So there is a lot of focus on the EDGE, basically the RAN infrastructure components, which I'm going through. That is where, actually, the RAN, Open RAN ecosystem is focusing on.

And then there is this hyper personalization of modulation standards. This one is basically what we also call as the cognitive RAN. This is one of the very interesting spaces wherein actually even Sanjay is working as well. This is a thing where we say that actually, you could be modifying the standards dynamically, right?

Unlike the current standards, where you basically have a 4G or 5G, and only some schemes of the MCS changes. You could potentially define standards on the fly with this kind of a communication systems.

And then with the Open RAN coming into play, actually, there is a lot of transition of the infrastructure elements to EDGE, and also some of these going towards the COTS based platforms actually. So that's one of the key things that is actually happening as we speak-- that actually, moving to a COTS platform actually gives the users tremendous power, in terms of actually being able to operationalize and even manage their own networks, and value add because they are the ones who know what exactly the network is.

So moving to a COTS based platform actually all of a sudden gives the people an entirely new paradigm, in terms of being able to have their own set of people actually manage, monitor, and modify the network to their needs.

And the next part is the Java-fication of the RANs leads to the RAN as a platform. Like, we all have heard a software as a service and things like that. So this RAN is a platform, is an opportunity wherein actually you could have service providers who can actually rent out spaces on their premises to run specific RAN elements. Moving to a card service platform actually gives us that particular edge as well.

So while these are things related to the virtualization, that means that taking away some parts of actually the core RAN ecosystem, and then moving towards a COTS based platform, and specifying how exactly you can have your partitions wherein you can run things on the edge, as well as on the cloud kind of thing, right?

Then there's actually the other part, is actually the heterogeneous traffic flows. So we have things as varied as actually IoT tweets, which are actually a few characters to enhance mobile broadcast systems, like eMBB standards kind of a thing. So here actually, we are talking about very high throughput of greater rate requirements.

So actually, the wireless ecosystem actually has to support a plethora of data rate requirements. So there's a tremendous focus on actually ensuring that everything starting from a very, very low data rate thing up to a few mbps of transmission, everything comes under the ambit of actually a unified standard, basically.

And IoT traffic and devices also requires a platform centric approach. Otherwise, once again, it's like, it's very disintegrated otherwise, right? So if you don't have a common modality of communication, then everyone tends to implement their own standard, and actually the scale up and things like that becomes pretty difficult.

And by this, you also give an opportunity to develop a lot of cost efficient devices that could essentially be a key differentiator as well.

So the other part of actually the space where we are looking at, where the wireless technology is headed, is actually the co-existence of heterogeneous networks. So primarily, as what we have seen the terrestrial cable and satellite, these-- and cellular-- these have been quite disaggregated, in the sense there has been no common standard that's unifying these spaces.

So if you look at it actually, there was once a time when actually the satellite part of things, especially with reference to DTT kind of things, moved very, very fast in terms of actually delivering video broadcast content to direct to the home. These were satellites, actually.

But they were never actually unified with reference to the other technologies, and they are cellular, even though it started much later on, they leapt up quite nicely, and we have a lot of things with reference to high data through codes, everything working with the cellular.

So cellular is now more focused, in terms of actually the 3GPP community and things like that working towards a lot of standards and improving the standards and all. Satellite kind of got left behind. And with the next new releases, wherein the 3GPP is also looking at unifying these terrestrial satellite and cellular spaces, bringing them under the same ambit of wireless communication standards.

This will supposedly be good for a whole lot of reasons. But that's one, and then satellite IoT and hybrid terrestrial and satellite devices, once again, the same. So when we say unifying the satellite and cellular networks, it also is essentially bridging the heterogeneous traffic flows as well.

So DTT is one more attractive option for video offloads. So currently, with reference to the cellular architectures, we actually have-- even though there are certain standards, like EM, VMS and all of which are being proposed, and then being voted upon, still nothing has gone very concretely into this, even concretely into the standards, and also we haven't seen devices coming out there.

So there are several options also that are being targeted here, wherein actually there is some kind of a convergence that is happening between the broadcast and the cellular perspectives.

So now there is also alternate technologies. There are certain distinct nature, for India now, this thing wherein we have large remote rural areas which needs connectivity, and the cellular doesn't really fare well in those particular things. So 3GPP has the 3GPP announcements also. There are some works that are happening within the LPTV and the elements of low mobility, large coverage kind of a target market is also being played within that.

So these are some things which I feel are some of the areas in which some works are happening, pretty diligently as we speak. So we are set for very interesting times, as far as the wireless technology space is concerned.

So just to brief on what exactly are things that are happening in the 5G. So as we know, really, 16 is something that's already baseline, and a lot of work is happening on the release 17 and 18. And when we say the new releases, they are not restricted just to the RAN elements that has been focused, in terms of the system architecture, securities, management, orchestrations, because virtualization is one of the primary things that is getting in into the 3GPP as well.

So that also there-- there has been work items that have been stretching out on all these things. And a few radio layer enhancements, the RAN level enhancements are primarily in the MIMO, the CA carrier aggregation parts of it, and some positioning coverage announcements.

The list is pretty long. Just wanted to say that actually with every new releases, there is always a lot of work that is actually going on in the RAN space, right?

So this specifically brings us to the point that when we say RAN level enhancements, it means that we have a lot of algorithms and things to be simulated, analysis of the performance to be done, new algorithms to be formulated, and then reiterated again and again. And these are the ones that will feed an even into the specification aspects of things, as well as into the final product as well.

So then we have basically-- the thing is actually, with every new paradigm, the releases that comes in into the picture and all those things, there is, like as we briefly touched upon in the previous, this thing with IoTs and things like that coming in, we need a lot of cost effective solutions.

Given the UE chipsets, also needs to undergo certain changes, maybe with reference to the hardware as well. So there is a distinct 5G advancements, actually brings with it a very tight coupling of new wireless chipsets also that comes in to play.

And these releases of 3GPP are not just limited to UEs. They are, in fact, the entire Infra, as well as the UE elements, which are being portable. So the current trend is also to actually bring in a lot of Infra elements also as chipsets. Like, it's no longer just this thing of actually running just as a COTS, but even being able to use a very low grade COTS with specialized chipsets in order to make the Infra components also work low on power and be more agile.

So a typical algorithm development cycle whenever a new chipset or a new algorithm needs to be designed, essentially on the wireless space, you basically have your transmitter and receiver section. So you would look at some standard, look at the specific standard. Or if there is something that you need to contribute to the standard, you probably have your concept well in mind, and then you go and do some implementations, validations.

And then, you iterate basically on your receiver performance setup. You include the required channel model, and then yeah. This is the typical cycle, what actually happens as you look at the algorithm development is concerned.

And now, the thing is whenever we look at a wireless chip design flaw, it's not necessarily wireless chip design. These are some common elements that goes on into any kind of a chip design. I'm sure Muthurkrishan will add more value to this because he has seen through this umpteen number of times.

So basically, we start with the requirements, and then we do some required hardward/software partitioning. And then actually, there is no chip that is there without software. So a lot of you will see a lot of back and forth between the hardware and the software, even during the definition stages of the chips. As the software comes in, actually, there will be intermediate functional verification, wherein the software base will put the software that is developed onto RT simulations, or on FPGAs to do verification.

And then the rest of the things, like the synthesis place en route to get to the final chip shape, chip takes place actually. So this is a typical wireless chip design flaw. And the requirement of the algorithms and these things actually exists on both sides of the spectrum, where the hardware as well as the software will actually deal with the algorithmic flaws.

So while we look at the possibility that a lot of these interesting things are happening at all, nothing is there without a good set of tools. So for realizing a semiconductor, you require a plethora of tools. Similarly, to arrive at good algorithms and algorithm designs performing to the required specifications, being able to simulate it, well, you require tools like MATLAB.

I think Saankhya has been very sim-- very important much user of MATLAB for a very, very long time now. We even follow our current 5G space where we work on the use of 5G and the toolboxes from MATLAB, along with the communications. And a lot of RF related stuff also is something that we use, and we also use MATLAB Coder.

So pretty much, there is a lot of advantages of actually being able to use MATLAB kind of toolboxes. It gives us really quick time to market the simulator, and then visualize it better. We understand where exactly the problems are, and then we are able to iterate on the algorithms also much faster.

So yeah, in summary, I would say we are looking at very interesting times, as well as the wireless space is concerned. New technologies and new standards coming in, new releases coming in, new chipsets being designed in the wireless space, which is looking at actually ensuring that even the infrastructure markets are becoming quite important, even from a small company perspective, wherein people can add in values, even in the infrastructure space.

And similarly, on the UE side also, there are a lot of things that is happening, right?

So all these things, whether we are looking at it as an pure play enbale developer or actually, and chipset developer, some of the tools like MATLAB and all play a very important role in actually getting you to the market quickly.

So yeah, with that, I just conclude my small talk. Thank you.

Good evening, everyone. So my name is Muthu. I'm the CEO of SFAL. And first of all, thanks to Professor Sanjay Gopinath Maheen, and everyone from MathWorks team for giving us an opportunity to talk to the audience. We don't know the profile of the audience, but thanks, Sunil from Saankhya Labs for setting some keynote kind of address.

Even though a lot of very heavy topic for most people, I would like to talk about SFAL because that is what we have been instructed. Before that, I think a little bit about the wireless and radio space, I think a few questions came regarding what is RAN.

So I think the radio actually means manipulating electromagnetic waves, using kind of waveforms. We send some data, the codified on a carrier level, and on the other side, you have to decode and take the message carried from the carrier. So there are multiple ways of doing it.

So I think companies like Saankhya, people like Sunil and all, are handling this in many, many different ways because the communication, all of you know that, right from Marconi, when he invented radio, it has been there from maybe hundreds of years.

So there are lots of techniques deployed in effectively communicating through the air. So those are all the Radio Access Network, they call it. And most of you know that we create more work for ourselves, or we create a lot of software layers to define so many protocols and methods so that we communicate very efficiently.

Right from language is where we define many constructs, so in radio communication also, there are a lot of constructs. I think maybe some young people listening to us can go and figure out how many jargons were used in Sunil's presentation and go and figure out each one of them. Probably you can write a small chapter on each one of them.

But it's a good thing to just take it as a challenge. Go and read about a few things, and it will be useful.

So coming back to SFAL. SFAL is a partner to the government funded, and India Electronics Semiconductor Association initiative to help semiconductor and electronics startups in the state, as well as in the country. So it has been operational for close to two years now, two plus years, so it is also called a center of excellence for Fabless startups.

So this was-- like every initiative, we have to identify a problem. This may be an interesting slide for most of you because from the technology keynote, I would like to give a perspective of the ecosystem, as well as a business kind of a perspective. And that's where we really very happy to be associated with networks who helps our ecosystem companies do a lot of work, which Sunil mentioned that his company does for many years.

So what we identified was depicted in this right-hand side. It's a funnel. If you look at it, we only have less than 100 such companies. That's why if you see, we are at the bottom of the funnel. If you go and figure out within that 100, there are about 30, 40 companies in our state, which is Karnataka.

So we have depicted this funnel. For example, in China we have close to 1,000 such startups. In Japan, Korea, Taiwan put together as an own region, we are close to little less than 2,000. If you take Europe and Israel-- usually Israel is combined with Europe-- it's about roughly 2,000 plus. And then USA, where maximum innovation activity takes place, especially in deep tech areas like semiconductors, that are about more than 2,000.

So there is a reason we have defined this like this, because if you look at it, it is divided into five levels. We call it the levels of maturity. That means a company in India, this is probably looked at as a maturity level one. I think it's even about the same level of performance. People would like market, it has not so much mature if compared to a tech startup in the USA. But these are all the opportunity for us for India to go and bridge these gaps.

So maturity wise, we see that there is a 5x opportunity to go up in the value chain. But in absolute numbers, we are looking at about 20x. That means 100 to 2,000. So, 20x.

So the integral opportunity is the multiplication of these two. It is kind of convoluting these numbers with this, one of the common modulation, demodulation technique is called convolutional-- deconvolution, even in the communication world, which is what Sunil was alluding to.

Even for data analysis, we use the same thing. So our integral opportunity is 100x.

So this was the problem identified by the industry people, and they took it to the government of Karnataka and convinced the government of Karnataka to give some funding. So that how do we go on, reverse this pyramid and normalize this pyramid.

So with that objective, we took some money from Karnatake and applied as an infrastructure, and tried to create more startups. And you can see the results in the left-side table. The left-side table is the result we have achieved so far.

We set a target of 50 companies in five years, and 12 products coming out of those 50 companies to be operational for a little more than two plus years now. At just two plus years, in real sense. So we have selected about 23 companies, out of which nine of them are working very closely in building products.

And there are a lot of due to pandemic, as well as lack of capital, or due to priorities that a lot of companies are still in discussions, even though they are selected but they are not started. Two companies basically died because of some reasons. So many of them are not able to raise enough money. They're not able to assemble the best team.

So 23 are still basically active, are in discussion on designing products out of 25. So this 25 is selected from the bottom 100 in our funnel.

Apart from this, we also help the budding companies to complete the small-- these projects run for 18 months. We have an engagement for days time of nine months, followed by three months durations, which are reviewed and extended every time the review becomes positive. Apart from that, we also support all the companies who approach us. Did some short time cohort program, we call it. So we have completed about 11 such cohorts.

And now that I report 20 plus companies are waiting in the pipeline, we go through an industry expert panel selection process before we engage with those 20 plus companies in the pipeline.

How do we accomplish this? I will explain it in the next slide using a partner's network image map. It's one of the key partners. That's why we were invited for this webinar also.

And we also have a lot of mentors to support our companies with respect to anything. These mentors can solve some of the problems that the entrepreneur are not able to create some ideas, even from hiring people, finding some technology solutions, business enablement, fundraising. All those activities, these mentors will help. And we have created a network of about 40 plus mentors to help our companies.

And more importantly, we help the companies to raise money directly. So far, amongst these nine companies which are actively involved in projects, some of them-- a few of them have raised close to 18 crore rupees as capital to help them.

So now these companies are into creating about 8 plus products, and there are a lot of patents under each one of these product line. We have one company, which has filed about six patents in photonics, for example. But there's a new upcoming field, using light to communicate data. Literally removing the copper and aluminum wires from computer to computer in data centers.

So there are companies like that. They have filed about six to seven patents. Like, Saankhya also has to pretty close to 100 patents in their last 10 years of their life. This company just started, and they have already filed six to eight patents.

So these companies create products, so that's why that's a whole motive, so we are halfway through the life. We have already created impact on 25 companies and 8 products, which is somewhere exactly the midpoint of the target set for five years.

And by the way, we also very closely-- some of you may know-- that government of India have come up with the massive semiconductor policy of $10 billion, which is 76,000 crores support. And after that, we see a lot of interest within the country, as well as outside the country. There are a lot of people approaching to India to do a lot of activities. So we will be facilitating such investment inflow to the country.

So this is how we accomplish. So far, we have accomplished the results. Our ecosystem is extremely strong now. So we are in the middle. The SFAL is in the middle. If you see the right side, you can see the MathWorks is star marked on the right, as you can see. So they are one of our key partners.

They really give a lot of benefits to our ecosystem companies, that are close to six or seven companies, today benefiting from access to the tools, which varies from basic algorithmic simulation to a lot of domain analysis capabilities and modules.

And we have access to earlier design tools. We have access to foundries. We have access to FPGA partners. We have access to embedded ASIC directions, and we have four selected activity partners also.

So we have-- almost every week or every month, we add a few new partners. So that's how we create this ecosystem, which helps all our companies.

And this is not about the domain. So we have few other partnerships also. For example, we help companies raise funds by working with fund partners. For example, if you see all this left side, the funds are all working with us. And they look at our companies, and then they evaluate them. And if they find suitable for their investment thesis, then they go ahead with investments. So that's where we could raise the data in growth for all these companies from these funds.

And we keep adding. There is a corporation called the CVC corporation, one of the big corporate company out of Japan. They also have a big office in India, so they are very, very keen to work with Indian companies by investing, by taking them, their business to Japan and other places globally.

Not only that, we have-- government of Karnataka has another very interesting initiative called the Global Innovation Alliance, which is very closely associated with global corporates. The reason is, the global corporates will be very helpful to take Indian companies to abroad as a partner strategic partner. So we are working with Applied Materials, three of our startups are engaged with them.

And along with their NXP, the every year contact the Startup Challenge program. Last year, we have got 10 startups through it, and it is interesting for all of you to note that three of such startups went and raised money in the famous Shark Tank program on the Sony TV. So three of those companies actually belong to that batch.

And right now, NXP is running-- the applications have started, and by March 18th, they're closing their application for their Technology Challenge 2 program. So that is why NXP.

And Qualcomm is also we are working with very closely to help our startups. And a lot of other companies like LDI, Imagination, all of them are talking to us to support the Indian ecosystem. I think this is really the moonshot idea for India, this decade is going to be communication, semiconductors, VLSI, embedded systems, and electronics. So I think almost everybody who is attending this webinar someday will be connected to some of these elements, and we are all in very good field today. This is going to grow bigger.

And not only this, the other operations partners also, even though they're all support functions, we have partners for them. We have space partner. We have ID PDA license management partner, according partners, legal partners. They all work with our companies.

And we also expanded our geographic profile. We are operating out of MG Road in Bangalore. Now government of Karnataka has given us exclusive space in layout. We have expanded to that also. So that shows that our ecosystem is growing physically also.

So these are all the companies, I think. In fact, Sunil's Saankhya Labs is also one of the companies because we don't just work with the startup, we also work with the grown-ups late stage companies like Saankhya, Signalchip, who are all into 4G, 5G chip design.

So they also can work with SFAL. They can come to SFAL to create next product. So that's how MSMEs are supported. Apart from the early stage startups, the rest of them are all early stage startup. But Calligo Technologies is one of the promising high performance computing, the boship number system based compute engine they are building.

Saigeware stage where it's very, very interesting health care, AAM hardware model based company. Kalatronics is building a high speed set of IPs. Lightspeed is probably the star portfolio company for us. They are building photonics interconnects using light to communicate the data.

In fact, as we speak today, they won landscape award in both USA and Singapore this week actually. So we have another company called Chip Spirit, which is into cybersecurity for defense applications. We have ABCRL, which is actually doing a tire pressure monitor sensor associate, and which is also very, very unique product.

And Morphing Machines is a ISE incubated startup, which is into reconfigurable compute. I think Sunil talked about how the domain specific architectures are architected, whereas Morphing Machines is building reconfigurable DSEs actually.

And there was a woman picture put next to it, just to indicate that it is led by Dr. Ranjani, a woman, so that we track to promote the women in deep tech areas.

So out of the selected companies, the 14 of them are still under cohort, so that is one phase before the selection and they get into some kind of a deep tech engagement with SFAL, so they are 14 of them. And we have, as I mentioned earlier, we have completed about 11 such separate cohorts. And there are 20 plus companies are waiting for us.

So I think some of you may know about the government of India's design linked incentive and FAB schemes. So SFAL has been doing similar things, but not necessarily FAB, but in the startup ecosystem. So we would even be happy to say that we are a precursor to that scheme.

So this is my last slide, for giving me 5 to 10 minutes. So it's very simple. You can go to SFAL website, www.SFALCOE.com. It's very simple, five minutes application. And if you go and look at the apply to SFAL, and you can fill the form, if anybody wanting to start a small startup.

We always encourage people working in the industry to come up with an idea, and convert that idea into some kind of a product. I think SFAL is the best place to come.

And we also wanted to highlight that the Register button on the right-hand side, if you notice, if you press that-- if you go to the website, you will be directed to the Becoming a Member IESA, as well as a registering a entity in start up India, as well as register you entity in start up Karnataka.

So I think I would like to end here, but I would like to re-emphasize the support of companies like MathWorks, and it's an anonymous support. They give almost like entire toolchain of networks is given to our companies for one year. And this is part of their Global Incubator Accelerator program.

Not only that, I think they're always available to support with the documentation, technical documentation, or creating webinars like this, to make sure that our knowledge level improves.

So we thank MathWorks for giving us an opportunity. But I would like to emphasize that if any of you or your friends in the company, outside the company, is interested in doing anything in electronics realizes, semiconductors, communication, spanning any field, I think SFAL is the place to come. Thank you very much.

So thanks Mr. Sunil and Mr. sumot for the great introduction. So it was very insightful. And going forward, thanks thanks, everyone else for joining this call today. And it was nice to have you all here.

So going forward in this talk, we will see about various wireless connectivity scenarios, and modeling various scenarios for the wireless communication system. So we'll deep dive into various aspects of the technology, which is available, and then study about the various way in which it can simulate this one.

So I'm Uvaraj Natarajan, and I'm from the application engineering team. Along with me, I have my colleague Tabrez Khan. So we are going to take you to an exciting journey of various technologies, which you can simulate the wireless scenario as such.

So I have my agenda something like this. We'll start with an introduction. And we'll go over 3GPP based 5G and LTE based system, and then WLAN. And then we will also jump on to the space area satellite communication system. And then coming back to the ground, smart home IoT standards. And then something will be lots of talk about the hardware deployment, and how you take the system, which you are designed into the hardware platform, and then prototype things. And the so on.

So we'll also take the questions at the end of the talk. I would request everyone to post your questions in the Q&A or the chat window. We will take up all the questions at the end of this talk.

So to give you an introduction about various wireless connectivity that is available, the 3GPP based 5G, LTE, and 3G plays a very important role in the mobile networks. So they define how the mobile network should operate, and how it has to be designed and so on.

But we also have various other technologies and standards, which is quite different from the 3GPP based 5G and LTE system. But it has its own characteristics. And they are designed for a different kind of purpose, other than the mobile communication network. Like, they are designed to work on an unlicensed spectrum, with a different deployment characteristics, an ad hoc network, or a mesh network, or even indoor scenarios, which work on very low and low power and low cost and very high efficient systems, which operate with better battery life than the mobile communication system.

So all these things put together form the complete wireless communication ecosystem. And I have various standards, which I have put on below, which works on the wireless communication system as such.

And implementing this standard based specification has its own challenges. So designing a proprietary wireless communication system is something different, but implementing a standard based wireless communication system need you to, or need an engineer to understand and read-- read and understand the specification thoroughly.

And first of all, you should understand the theory, and then understand the practicalities behind that before you jump into the implementation. And then the next step for him is to develop the algorithm message, which is based out of this standard, which is defined that you need to write the code, develop the algorithm, and then test the algorithm, which he has developed. And the thorough testing has to be done in order to make sure that it complies to the standard set.

So all these are challenges for an engineer, and the standard compliant Toolbox from MATLAB are the various toolboxes, which we are going to touch upon today, will help you to solve these problems easily because there are various things which are already implemented for you to make your job simple.

On that note, there are various applications and use-cases, which I have bucketed into, which will help you to understand the wireless ecosystem message. So when you start working on the wireless communication system, you need the wireless waveform.

So the waveform forms the first and the integral part of the complete system, so you need to generate a standard specific waveform, whether it's a 5G waveform, or an LTE waveform, or a WLAN-based waveform, or even a satellite communication waveform for that instance. So you need a wireless waveform which has to be generated.

And then you need to be able to simulate the system into it. So it will be a link level simulation or a system level simulation. So system level simulation, I mean by having multiple transmitters. For example, in 5G, 6G would be with multiple UEs, or multiple loads with multiple receivers and so on. And then see how the system behaves as such.

So that's the next step which you will have, the simulator. And once you're done with the link and system level simulation, you will have to do various measurements to see how your algorithm work actually. So these measurements are in the transmitter, and mostly on the receiver side, so you measure the EVM, ACLR, and various other measurements like spectral mask and so on in order to make sure your algorithms are fine.

But once you're done with all these simulation aspects, you take it to the radio. So you connect your hardware, either it be RF signal generating or capturing equipment, or the software defined radios, and then do over the air kind of thing.

So this forms the complete lifecycle of developing any wireless communication based system. And our customer Lekha Wireless have used MATLAB and various toolboxes in order to perform unit level and functional level testing development, and also performance validation of various algorithms of the complete system using the MATLAB 5G Toolbox and various other toolboxes.

So with this here, they were able to remote cut down the development time, as well as testing time, so that the time to market has been reduced a lot in order to keep up with the competitive environment.

So we will have all these slides mailed for you after finishing this talk, so you will get a copy of the slides as well as recording for you to take it for.

Let's quickly get start with the 5G and LTE based system. So 5G communication, as we saw in the previous talks as well, have a lot of highlights which has to be concentrated when we start developing a complete data system. So again, 5G waveform can be generated using the 5G Toolbox, which is a part of the MathWorks.

Along with that, you have various other capabilities, like generating the channel model, generating the complete into and link level as well as system level communication as well, along with the throughput measurement, and then synchronization procedures, and various other parts of the system for the testing and measurement can also be simulated as a part of the 5G Toolbox infrastructure.

So with that, I have a very interesting application to show you, which is Wireless Waveform Generator App. So using this application, you'll be able to generate the waveform. Any kind of waveform, like the 5G, LTE, WLAN, Bluetooth, any kind of waveform can be easily generated, and the good part that is you'll also be able to generate a custom waveform which is actually a customized waveform master requirement. And then generate the waveform easily with the impairments and exported into your system, so you get a ready made waveform using this application.

And once you have the waveform, you'll also be directly transmit the waveform in the air so you can connect your instruments to the instrument control Toolbox directly into the Wireless Waveform Generator Application. And then you see your instrument there, and then you can directly transmit the waveform, which is generated, over the air directly on the transmit button.

So we support various instruments, either it be hardware-based instrument by third party hardware, or it can be an SDR like Pluto or USP or Zynq-Based SDRs and that's available for all the technologies, like 5G, LTE, WLAN, Bluetooth, OFDM, and so on.

Then going to link level simulation-- so when you are simulating the complete system, channel model plays a very important role. And there are various channel models which you should consider, depending on your applications. For example, starting with the basic AWG and the fading channel, so which is purely stochastic, or you'll be able to generate the geometric or ray tracing based channel model.

So 3GPP also defined various channel models based out of LTE and 5G standard. For example, 5G defines the TDL and CDL, cluster delay and tapped delay line, which actually mimics the 5G based communication system in a simulation environment, and you will be able to insert the channel models into your link level simulation directly in the 5G Toolbox in MATLAB.

And when you're working on a beamforming based simulation, so you should be interested in MIMO based channel model, right? So you should be able to simulate a MIMO based system, and that can be, again, done directly using a MIMO based channel model, which is available. And then compute the free coding matrices, and transmit the signal, and then receive it, and see how the signal behaves, and so on, with the help of various other functionalities like the ray tracing and then so on.

On top of this, you'll also be able to download a part of the map from the internet, or maybe like OpenStreetMap, and then imported into MATLAB environment. Place your transmitter and the receive antennas in different locations, and then connect this environment into your simulation environment so your 5G runs, 5G system runs, waveform is transmitted through this antenna, and here you nicely see a visualization where you will have your coverage analysis of the complete system.

You'll be able to steer the beam, and then take it forward from one user to the other user. And then see how the performance of your system of your algorithm behaves and so on. So these are some nice visualizations, which are again, easily simulated using MATLAB.

Coming to the system level simulation, which is actually one step above the link level simulation. So we have certain part of the protocol stack, the 5G protocol stack implemented as a part of the 5G Toolbox, where you have the capability to simulate the complete physical layer. Along with that, you also have that MAC layer, RLC, and then various traffic generation, like the On-Off traffic, FTP traffic, or even voice based traffic can be easily generated.

So you also have the flexibility to design your own scheduler using the 5G Toolbox, or plug in any of the schedules which is already available in the system. And then you'll also be able to simulate a multi-cells scenario, that you have a single cell with UEs, and then placing another cell next to that, model a quad channel interference kind of a scenario between multiple cells directly inside the simulation environment.

And once the simulation is done, you'll also be able to log the packets in the form of PCAP, or input the PCAP from your other simulations and then do analysis on top of it. So this kind of 5G based system simulation is also easily doable in the 5G Toolbox.

AI for wireless communication is, again, another important aspect, which we are focusing on and which we are highly investing on. So there are various scenarios which has to be supported when you work on wireless communication related problem statement.

So what is the AI for wireless communication means? It means using an artificial intelligence based technique, like deep learning or machine learning or reinforcement learning or Federated learning, any of those techniques to solve the traditional wireless communication related problems.

So there are various traditional problems, like the signal classifications, or RF digital pre-distortion algorithms, the receiver algorithms, beamforming, encoders, the decoders, channel modeling. So these are common wireless communication related environment, and you'll be able to easily combine or develop deep learning based system, and train the network, create a network, train the network, and deploy the network in the MATLAB environment to solve the same problem statements thanks to multidisciplinary workflow, which is available in MATLAB.

Like, you have support for deep learning and machine learning in the same environment, so you'll be able to combine the wireless technology with the deep learning technology and work on it in a single platform.

So we feel that the time for AI is coming now because the service is expected from this system in the traditional way, and that services needs to be scaled up like anything. And as the data rate increases when we talk about the 5G and beyond 5G. So the traditional algorithms have to be rethinked off, and so that is where the AI is coming into picture.

What if we use AI based system, rather than using a traditional algorithm? And we know that the AI shows a very promising result in other domains as well. AI is not something new, so it has been there for years. And so we have implemented various algorithms, are or various ways in which AI can contribute to the system, like the network simplification or improving the performance of network or spectral efficiency.

So as a part of MATLAB suite, we also support solving wireless communication related problems using AI, so we have various shipping examples available, like the 5G channel estimation, or spectrum sensing using deep learning to identify the 5G or LTE signals, or even the deep learning based communication, or beamforming using deep learning. So all of these are various new areas which we are working on.