Smart Factory Summit

Welcome to the Smart Factory Summit. MathWorks and FSM will be hosting this event to showcase technology and workflows that will help you go through the Fourth Industrial Revolution. FSM, the foundation for smart manufacturing, is an initiative at the Indian Institute of Technology, Delhi, with collaboration from the Automation Industry Association and the Indian Ministry of Heavy Industries and Public Enterprises.

FSM's charter is to promote the adoption of Industry 4.0 solutions to the Indian manufacturing sector. To begin this event I welcome Mr. Prashant Rao, who heads the Application Engineering Team at MathWorks India. And it is over to you, Proshant.

Good morning and good evening, good day to everybody, all the attendees of this session. As you heard, my name is Prashant Rao. I head Application Engineering for MathWorks in India, and I extend a warm welcome to you to the Smart Factory Summit. This summit is a culmination of two different things, actually. One is the start of an engagement that we have got.

We've actually been partners with FSM since mid or the latter part of 2020, but due to the pandemic issues, we weren't able to do a physical event to announce that partnership. So we're using this opportunity to do that on the one side. This also has given us some time to prepare some technical demonstrations that we can talk about today as well.

The other part of this is that it's a culmination of a series of webinars that we have had over the past few weeks around smart factory and enabling AI in smart factories, so we had three parts already, three sessions already of this, and in case you were not able to attend any of these and you are interested as an attendee to take part in this, we have the recordings for these sessions available.

So please feel free-- I'm leaving this slide up for a few seconds-- to scan this link, the QR link on the right-hand side, or look up Smart Factory Webinar Series on The Mathworks Home Page and you'll be able to find the other sessions as well. Today is a culmination of the series, in that sense, where we have the Smart Factory Summit, with a series of speakers.

We'll have Mr. Anup Wadhwa joining me in the introduction to the session. He's the Director of Automation Industry Association in India. We have Professor Sunil Jha, who will deliver the keynote address. Professor Sunil Jha is Director of the Foundation for Smart Manufacturing. He is also a Professor at IIT Delhi in the Mechanical Engineering Department.

We have Philip Wallner joining us from MathWorks in Germany. Philip is our Industrial Automation and Machinery Industry Marketing Manager, and he'll be sharing a few thoughts from his side as well. And last and absolutely not least is Mr. Shireesh. He's the Practice Head of Digital Transformation and Industry 4.0 at Tech Mahindra. We'll be sharing some insight on how he is looking at Industry 4.0 in the context of Tech Mahindra's work.

We have a technical demo team comprising of team members from FSM, Vinshu and Omkar, and a team from MathWorks as well, a few application engineers from the team. We have Peeyush, Ramanuja, Rahul, and Shripad, who will be working behind the scenes. Sometimes, you'll hear them, but they've done a lot of work behind the scenes as well. Right, moving on to the agenda for today.

We have a brief introduction note. After a few minutes, I'll hand it over to Mr. Anup Wadhwa a few thoughts and introduction. From his end, we will have a keynote address from Professor Sunil Jha. We will move to Philip Wallner, who will talk about the factory of the future, and then the industry presentation as mentioned earlier by Mr. Shireesh around the role of digital twins in 4.0, Tech Mahindra's point of view and experiences.

We have time at that point in time for a Q&A. Please post questions in the chat. We may be able to take them and curate them from the chat and answer them during that Q&A session as well. At times, we may answer the questions directly in the chat as well. We have a five-minute stretch break. We have two videos that will be running in the break, but this also gives you an opportunity to get some coffee, refresh yourself.

And then we have a technical demonstration session where we'll be showcasing through three different demonstrations how we can transform from a traditional to a smart factory setup. We have demonstrations around building a smart lathe machine, digital twin modeling of a machine and virtual commissioning with PLC. Integration. So let me spend a few minutes to talk about your hosts, specifically now with MathWorks.

MathWorks is a leader in the 2021 Gartner Magic Quadrant for Data Science and Machine Learning for the second year in a row, and this is on the basis of our completeness of vision and the ability to execute on this vision. The difference is that we understand that AI success demands more than just an effective model or an algorithm. It's about delivering a full product or a service to the market that is based on AI, and we approach this problem by enabling engineers and scientists to incorporate AI across their entire system design work.

A few more words about The MathWorks itself. You may be knowing some of our products. MATLAB is a programming environment for algorithmic development data analysis, visualization, and numeric computation. And Simulink is a graphical environment for designing, modeling, simulating, testing, verifying systems. The work that you do in MATLAB and Simulink can be deployed in any kind of system. It can be deployed on embedded systems, desktop systems, on the cloud, and there are many other options for that.

We have a whole set of 100-odd different products for specialized tasks. Here is an example of a product known as Computer Vision Toolbox that is specifically for computer vision tasks. And these products span several domains across engineering and science.

Talking a bit about digital transformation in manufacturing, so we hear this term a lot, digital transformation. And the way we understand it in this manner is it's about the data integration between assets, edge, OT-- operational technology-- and IT systems. So whereas these systems used to be encapsulated in their own silos earlier on, now with the technology that we have at hand and with communication technology as well as technology that enables us sync between these different systems, we can actually send data seamlessly across the entire set of systems over here, these family of systems.

What does this enable us to do? So when you think about it, of course, on the one side, we can have integrated operations. So that means we know at the IT level, or at the organizational level, about processes that are happening in a specific asset or in a specific production environment. But the feedback loop isn't also, or-- possibly of more value. We can actually then optimize and manage processes at the factory level or at the asset level, but across the entire organization.

So optimizations can occur then throughout the organization, taking inputs and data from across the organization to enable that optimization. And finally, in when applied completely, this allows organizations to use business insight to drive and optimize manufacturing processes. This can also be used vice versa. You can use manufacturing processes to then optimize your business decisions as well. So overall, organizations find great value in being able to enable digital transformation across the workforce.

What does it mean for us when we're looking at it from a domain perspective, from a domain expert perspective, especially when we're talking about engineers and scientists? We believe that engineers and scientists should be enabled to do the work that they are demanded to do in the best possible manner. And we provide tools that enable domain experts to do their best work. Now moving from traditional- or first-principles-based approaches, for example, from system design and development-- now incorporating also some of the newer data driven techniques, big data, AI, analytics, and cloud integration-- our tools enable domains experts to leverage the tools that-- or the AI-based applications or AI-based technology, across their entire workflow.

In terms of the partnership of what we are expecting or what we also do with FSM, we think that there are three key pillars to this. Basically, we would like to enable and accelerate the adoption of smart manufacturing methodologies in the market, and we can do this by focusing on technologies themselves, which is what you will see part of today. But also, by building competency, helping people to adopt these methodologies using the domain expertise, empowerment, collaboration.

So we are focused on that as well through sessions like this and also help organizations set up the processes that they would need to do be able to develop smart manufacturing or leverage smart manufacturing techniques. So whether it is setting up processes for model development, test validation, incorporation of AI, or these technologies across an entire enterprise and basically enabling the entire digital thread, we will be very, very happy to be able to help organizations do this.

I will hand it over to Mr. Anup Wadhwa for the rest of the introduction. Anup Wadhwa is an alumnus of IIT Delhi. He served in systems engineering and special projects for BHEL and Rockwell Automation as well as SAMTEL. As director of AIA, he conceptualized and steered the industry partnership with IIT Delhi and the Department of Heavy Industries to set up the IITD-AIA Foundation for Smart Manufacturing, FSF.

Anup G. is an evangelist of the Samarth Udyog Ecosystem that integrates the efforts of different agencies in government and academia. He has served on Government and UNIDO Expert Committees related to country-relevant initiatives and Industry 4.0 with special focus on interdisciplinary cooperation. Anup G., I'll hand it over to you to take us through the rest of this introduction. Thank you very much.

Thank you, Prashant, for that introduction. And also, thank you, Sunita, for sharing a brief on the IIT-Delhi-AIA Foundation for Smart Manufacturing. First of all, I would like to place on record my sincere thanks to you, Prashant, and the entire team of The MathWorks for being a very active participant-- I would say more a partnership approach, which your team has demonstrated-- and I'm very sure that you know, we would take this forward with greater speed now that we are getting ready for physical engagements.

Having said that, I would like to share with our audience a little bit about the genesis of Samarth Udyog and our foundation as a hub for Samarth Udyog. A few years back, Prime Minister Modi had named the chief guest at the Hanover Fair and India was also a partner country. And that was the time he was taken to a booth and shown a bit about Industry 4.0. And he was very fascinated and he foresaw that a country's economic development was very closely linked with how we would absorb and implement this new technology.

So he came back and he said things, you know, floating about. Who are the agencies? What can they do? How can India ramp up absorption in this technology? In our meetings-- the Automation Industry Association-- we began to discuss the implications of Industry 4.0, and we had several players from Germany, America, and Japan who were on our board. And you know, they had a lot of things to share about how their respective parent organizations were at the forefront.

One message was very, very clear that there was no single company which was claiming to be at the leadership position, and there were a lot of cooperative efforts which were going on in these countries to establish the foundation for Industry 4.0. I'd like to draw the attention of the participants to this slide. As you can see that there are two colors splashed. One is the blocks with green, and the other are those darker blocks.

Now, if you look at the blocks in green, they represent the established entities of an automation or an automated manufacturing ecosystem. So you have the automation players, you have the engineering process contractors, you have the machine builders, the line builders, control integrators, panel builders, distributors. This group of players were, at that time, already adopting well-documented engineering practices. We're very well aware of global standards and buyers, consultants were also pretty clear about how to tackle or chase and testing of physical assets, including the automation system.

So that was one established ecosystem. On the other hand, if you see the dark blocks, there was this emergence of the software-based industry players. So we had the industrial IoT solution providers, we had the cybersecurity providers, you had the IT Infra providers, and generally, we had a great buzz about cloud computing and internet-based solutions as the next big thing happening for the industrial world.

Now the dilemma was not about whether or not the IT and OT systems would integrate. The key concern was how do we, as a nation of industry practitioners, of educators, of skill providers, of standard agencies, how would we understand, absorb, educate, and implement going forward industry for solutions? So with that challenge, the Indian government took a leadership position and-- can I have the next slide-- we all realized that even the advanced countries had got a mission statement-- they had a flag-- and they were putting all their energy in a very well-coordinated manner.

So industry 4.0 was obviously the pioneer from Germany, but then you had China, you had the Americans with the AMP, and you had the Indian government giving a new twist to our mission, calling it Samarth Udyog. That was the genesis of the foundation for smart manufacturing, and I'll talk about that in a minute. Can I have the next slide?

So players from automation industry, consulting, and the software industry came together. And along with IIT-Delhi's gracious hosting of facility, we looked at several facilities. You can see the slide. We have the IIOT systems. We have the collaborative robots, smart sensors, rapid prototyping, AR, simulation, remote maintenance, safety, cybersecurity. A lot of this is a work in progress, but all these pieces are being knitted very well by Professor Sunil Jha, who is the chief guest for today's function.

You will get to hear about him, how he's going about actually building this facility and creating an infrastructure which is available for industry and academia. And as I hand over the baton back to Prashant, I'd like to extend a very warm invitation to all the guests on today's show. Please come to see this facility and help us in developing more and more use cases for you, for India, so that collectively, this can truly become a facility of national importance and relevance. Thank you so much and I also look forward to Professor Jha and the other esteemed speakers for their valuable contribution.

Thank you very much, Anup G., for your introduction. I will hand it back to Sunita as we move through the logistics of transferring prisoners to introduce Sunil Jha at this point in time.

Thank you Mr. Anup Wadhwa and Mr. Prashant Rao for this insightful opening address. I now welcome Professor Sunil Jha to deliver the keynote address. As director, IIT-Delhi-AIA Foundation for Smart Manufacturing, Dr. Charles is closely the manufacturing industry in absorbing new technologies to make them ready for the next Industrial Revolution. And it is overdue, Professor Jha.

Thank you so much. Very good morning and let me share my slide and then we can-- this. Hope my slides will be visible to you.

Yes sir.

OK. Thank you so much for this nice introduction in the start of the session today. And as you are already aware of this, is a series of the talks on the smart manufacturing being initiated by The MathWorks. And today, we are going to have this summit where we'll be discussing some of the methodologies of implementing smart manufacturing, the kind of solutions that we have developed, the kind of infrastructure and facility that we have created together.

So I will be just taking you through a very short glance on what kind of facility we have and what we are exactly doing, because in a short time of 20 minutes, it's very difficult to go a little deeper into the technologies. But I will give you at least the idea of the implementation of the smart manufacturing, and our foundation is certainly committed to bring out the solutions, develop the solutions, do research and development together with the leading industries in this domain, and then developing a solution, which are more adaptable by Indian industry.

So as we all heard about this, this is a common engineering facility center as briefed by Mr. Anup. That, as a part of this mission by government of India, IIT-Delhi has created this common engineering facility center with Automation Industry Association. So we bring together their strength, basically, what the strength lies with the industry. And as you all understand, we have partners now working with us all across the world.

These are MNCs which are basically contributing, sharing their ideas, thought process, the new technology developments with us. At the same time, as an academic institution, will also have the required specializations at IIT-Delhi which are certainly helping us in developing the solutions on the smart manufacturing. So with this initiative, and certainly the supported by the government and financially-supported supported by the industry and the government, we are able to build a world-class facility at IIT-Delhi to do research in this area and develop solutions for the industry.

So let me quickly brief you about the different stages in implementing smart manufacturing, because that's how we think about the implementation part. As we all understand, the basic foundation is to be built by the automation technologies. That's how we look forward, that the sensing, computation, and actuation has to be integrated. If they are legacy machines, if they are old machines, we need to integrate these basic ingredients, all the elements of the automation into the system. And post that, we would like to build the connectivity with these assets. And that is what the prerequisite for the implementation of the smart manufacturing.

On top of that, we can provide many solutions, many services that can be built upon. And the journey for the smart manufacturing implementation certainly start after they're done with. So there are multiple technologies that fall in these four blocks that have been shown on the right, how you can bring out the visibility in the organization, how the transparency can build across the different departments and the customers over here, how the predictive capacity of an organization can build up, and certainly we look forward for the more self-organized systems. The system should be able to adapt to the changes and the decision-making can happen. And that's the last stage of implementation of smart manufacturing is their adaptability.

So when we talk about the implementation across these different stages of implementation of smart manufacturing, we are talking about these key smart manufacturing technologies. And the idea, when we started this journey of understanding the smart manufacturing technologies-- certainly, we are familiar with the manufacturing domain and the automation technologies that are available as part of Industry 3.0 developments. And certainly, we look forward more cyber physical integration with the technology that are already available as part of Industry 3.0.

So we plan to do a lot of experimentation around these technologies. Certainly, you can understand these technologies well when you do experiment, when you understand these technologies better. So we created certain, basic infrastructure to learn these technologies and understand these technologies and create a platform where the people from academics, people from industry can come together and then do a lot of experimentation. Maybe these days, you heard about a lot of tinkering labs and the similar initiatives are done for the basic science learning and the STEM learning.

We can see that this is one of the infrastructure that is available for doing a lot of experiments in the domain of smart manufacturing. So we created this basic cyber physical assembly line in our lab, which is an assembly line for electromechanical components, how you can assemble the electromechanical component. Starting from the basic ordering process from the mobile application or the web application, you can just select the component to be assembled. You can click the order. You will get information, how much time it is going to complete the assembly operation and when your product will be ready to deliver.

So this assembly line is being created, mixing the technologies of the Industry 3.0 and 4.0 all put together in an initiative where the automation stations are being designed, developed, and integrated over here with the backend communication IIOT smart sensors being integrated. So whatever technology landscape that I've shown in the previous slide, they are all put together in this basic systems over here. And then, intentionally, we plan to put a heterogeneous hardware in place so that the similarity will be drawn in the manufacturing industries as well. Because there, you will find there are heterogeneous hardware in the automation systems to be operated together.

Then we have also added a state-of-the-art, multi-process robotics cell to the same facility. The idea is that on one side, we have more than 12 axes, 15 axes of the motions being executed by three different controllers with the three different HMI interfaces, how the networking is being done between the systems. And seeing now we have designed this multi-process robotic cell, we have the single, six degrees of freedom robot is integrated with automation technologies in place. So this small, three-by-three-meter footprint is basically capable of demonstrating the complete flexible assembly line where the robot can do the pick and place. And then the functional testing stations are being integrated as a part of this.

And starting from the basic raw material to the final and finished product being labeled and dispatched through this cell, that is the capability that is being demonstrated in this cyber physical assembly line. So we designed this multi-process robotics cell with the capability that it can do the assembly of multiple type of components put together. The orders can be customized, the routing of the product can be changed on the fly, and finally, a finished good can be delivered. There's a lot of learning happened as a part of this development in terms of how the integration of the robotics with automation can be done, how the IoT data can be used for controlling and monitoring the cell as a system.

So this is just a quick inside view of the facility that you can see there. So there's an incoming conveyor from which the raw material can be picked up by the robot. This is a ceiling-mount robot and which will do the pick and place for the assembly operation to be started. So we have assembly stations. Over here, we have a functional testing station, we have a nutrunner as well as the screen arrangement inside the cell itself.

We designed this system. Some of the indigenous development also happened as a part of this development activity, and we have this final packaging and the labeling system integrated into that. We have also implemented the capability how PLC can be used to control the robotics, motion control and the robotic programming, through one single interface. So that is also being implemented in this case.

Now we are also extending. This is one of the lab parts and we are extending the similar facility towards creating a smart factory, which is a kind of a micro production system, where we will be connecting together some of the CNC machines with the old machines, or old infrastructure like lathe machine, with a state-of-the-art robotic welding cell that we are designing and placing in place. All these are connected within a mobile robot, which is responsible for transporting the material, inspecting the material all across the floor in the smart factory.

We are also building some of the applications for the smart factory so that they can be utilized by the supervisors, the plant manager, the maintenance engineer to get real-time information about the production facility. So certain framework you need to be prepared how the data can be put together and how this information can be exchanged, keeping the data in premises and how the solutions can be delivered to the various stakeholders. We are designing all these applications for the smart factory.

Next development is basically creating a robotic welding cell. The idea is to demonstrate how our flexible manufacturing cell can be designed and created, which can be made available as a service because most of the time we see that the manufacturing infrastructure is lying idle if it is not being utilized for making your own products. So in that case, we are designing this cell so that it can be available as a service. You can have a web interface or the mobile interface through you can define your product requirements-- and certainly, in this case, a welding requirement-- how you will upload your data to the cell, and the cell will schedule the ordering, and then finally, will proceed for the welding operation, do the right quality testing.

So all is being integrated into this robotic building cell. And throughout the process, we will also be demonstrating how that automatic identification of a product can be done. Because each product that is being welded in this cell is a unique actually. So it's not on a repetitive welding operation as you see most of in the mass production activities. They are-- a particular job is to be welded in a particular fashion and whose data is being made available before the job is being submitted to the cell.

So we designed this architecture here, where the customer web application will capture all the required data through the cloud interface. And finally, the necessary data is being delivered to the local production server. And from there, the production activity starts. So the information is being passed to the master PLC because here we have multiple controllers to be coordinated among. So you have a master PLC that is coordinating the control between the building power source, the multiple robot controllers because this has two robots, one is the building robot, another is the material handling robot. At the same time, we'll also be integrating the welding inspection system inside the facility.

So this is a cell under creation. We are almost in the final finishing stage of this robotic welding cell. And certainly, as we allow the access to the visitors, you are most welcome to come and witness this facility. We have also created enough infrastructure for training because that's a very important part of our learning, that whatever time we spend in understanding these technologies last three years, we look forward the same infrastructure will be made available to all the learners from the industry and academia.

So we have this smart manufacturing training system designed and created as we understood how the learning can happen for the smart manufacturing technologies. These are made accessible remotely also, so that a lot of experiments can be done by the students and the people from industry, from a remote. They can try out their own concepts and implement on this facility.

We are also integrating multiple technologies, bringing out more demonstrations and the use cases over here. So that is how the smart manufacturing trainer kit integrated with the electromechanical hardware to demonstrate the various communication protocol in the industry, how the edge hardware where the vision processing can be done, and certainly, the kind of a toolbox-- Prashant was mentioning the computer vision toolbox-- all these solutions can be deployed onto this edge hardware. And then data can be exchanged with the OPC-UA, or with the industrial hardware.

So all these experiments can be done on this environment, and certainly, we can bring out the technologies like augmented reality to generate an augmented view of the information on the physical system. And you can certainly implement the check on the quality parameters with the various sensors. We are also working on the development of the solutions like an expert maintenance system where the multiple technologies are going to be put together in this environment and how we can build a system which can help us in identifying or analyzing the system behavior and finding out the fault. And once you are able to locate the fault, you can use the technologies like AR to locate and then do the replacement of the repairing part as a part of maintenance activity.

We are also working on integrating multiple solutions available around the robotics. That's how the robotic toolbox available from The MathWorks and the open-source ROS integration. All that is an ongoing process that we are working on. So putting an open-source technology along with the commercial technologies to make the solutions available and viable for the industry, that is the idea behind this implementation.

We are also integrating the digital twin part over here. And we'll try to build up a solutions like this where you can give the goals to the robot, and then the planning will be done with the open-source libraries, like move it over here. And then robotic motions can be controlled and the final solution can be deployed onto the robotic controllers and the extended controller.

We are also working on developing a solutions for the conventional machines. That is what we are going to demonstrate in the later part of this summit itself. Idea is to enable the same technologies, what we are using for the state-of-the-art equipments, how we can deploy the solutions for the legacy machines, and how we can convert that the whole infrastructure into a smart infrastructure in the facility. So we have already done that integration, the desired integration, in terms of the sensors and the controllers in place.

The idea is to monitor the product quality, control the product quality, on the manually-operated machines, how we can monitor the machine health. And finally, what kind of guidance can be given to an operator who is manually operating this machine. So that kind of solutions can be built, so that apart from controlling a machine through any manual control, we can also enable people to produce because most of our manufacturing facilities are having this kind of an infrastructure. So we will demonstrate that integration to you in the later part of this effort.

We are also working on the creation of the digital twin of the machine tools. That is the idea. So that is how the second demonstration is linked with that. So we have certain machines being created at IIT-Delhi. We are trying to build our digital twins of these machines. So idea is that how you can implement the simulations of the physical machine tools and how we can represent the machine tools with the different models. And then on the other side, you will be implementing the monitoring and control technologies together.

Idea is that once we have this fleet-based knowledge system put together, where you on one end you will have a simulation, on the other side, you have a physical machine running, and the control is being executed by the local controller, how these two systems will exchange the information between themselves. So how you will be managing the real machine data utilize to improve the accuracy of the simulation model? On the other side, you will have a simplified version of the simulation that is being used to monitor and control the machine activity. So that is how the digital twin implementation will be demonstrated in the next part.

We are also working on the training programs. That is how we are running the various training programs. So starting from a basic awareness through the deep dive on the technologies, how to do the implementation. So we have designed our own courses, which are running over a period of a month and a much deeper engagement in terms of learning the technologies and so on. So these are the training programs that FSM is conducting. So those who are interested can certainly join and then watch for the announcement in this directions.

Thanks all. I think that's what I have for you today. So idea is that we would like to demystify what all technologies being involved in the smart manufacturing. And we are open for all sorts of demonstrations, prototyping, research, all put together in this foundation. So thank you so much. Over to you, Sunita.

Thank you very much Professor Jha for this enlightening session on implementing smart manufacturing and especially the exciting sections on the cyber physical assembly line and factory. We wish FSM all the very best for supporting the industry in adopting smart manufacturing initiatives. Just a reminder for our audience, if you have any questions for our speakers during the event, please do put them into the Q&A box-- or click on the bottom right of the screen to bring up the Q&A box-- and our speakers will answer them shortly.

I now welcome Mr. Philipp Wallner, industry manager for industrial automation and machinery at The MathWorks, to take us through the session on the factory of the future. Welcome Mr. Philipp and it's over to you.

So Thank you Sunita Let me share from my end. So please let me know if you can see my screen.

Yes, Philip.

Thank you.

Thank you.

So hello and good morning. My name is Philipp Wallner, industry manager at The MathWorks. I'm working out of Munich, Germany and it's an honor for me to speak at today's Smart Factory Summit. The title of my presentation is The Factory of the Future. And in the next about 20 minutes, I want to talk about how models and the systematic use of models and data supports your digital transformation on the way to the factory of the future.

And-- well, in my talk today, I will do this with highlighting two specific practical examples of companies that are already on their digital transformation path and that are extensively using models and data for that. But first, let me give you a brief overview of The MathWorks and who we are. So while we heard a bit from Prashant Rao already, just to give you a bit more background. So The MathWorks has been founded in 1984. Since then, we have grown to a team of more than 5,000 employees who are supporting our customers worldwide. And more than half of our employees are developing our products and specifically, our core platforms MATLAB and Simulink.

For MATLAB, as we already heard, this used for developing data analytics, algorithms, and functionality for artificial intelligence. And then Simulink is a graphical environment for developing and testing mechatronics systems. And on top of these two platforms, we-- well, by now, have more than 100 tools for industry-specific workflows. So for instance, for code generation, for industrial controllers as we will see later on in my presentation, or for predictive maintenance, or for industrial communication through OP-CUA, as we just heard from Professor Jha. OPC-UA is pretty commonly used for connected systems in the manufacturing industry.

So our insights and our expertise on digital transformation comes from our work with many customers, many companies across multiple industries. And some of these industries come from a very electromechanical basis, electromechanical roots, and now shifting more and more into software-defined systems and more autonomous systems. Other industries don't have any physical products at all, so they are leaders in the area of using big data analytics and integrating with IT systems.

But all of these industries and companies in these industries are undergoing a digital transformation journey at the moment. And well of course, there are multiple differences between these industries, but we also see a lot of commonalities. So what does digital transformation really mean for the manufacturing industry and for the factory of the future? So what we keep hearing from customers in the manufacturing industry is the common theme of mass customization. So how can individualized goods be developed or be manufactured with the benefits of mass-- of the cost benefits of mass production? And this is actually really in the core of Industry 4.0.

I quoted a recent article from Rainer Brehm from Siemens, and he describes mass customization as producing what matters. So reducing waste by only producing what the market and the customer really requires and demands. So what does that mean for the factory of the future? So how can the factory of the future really be prepared for mass customization? So mass customization, out of our experience and our customers' experience, requires connected, flexible, and autonomous production.

And that means that the systems that are used in the factory are becoming more capable, they are more software-driven in order to be able to quickly adapt to new situations, to new products, to new parameters. And that actually means that these systems themselves-- the machines, the equipment themselves-- have to be connected, flexible, and autonomous. That sounds good but also comes with some challenges.

So recent studies show that the growing design complexity is among the top challenges that engineering teams in the manufacturing industry see. And if you look at the right hand side, this study from McKinsey, this is specifically true for the software parts, for their software, the machine application that runs on the machine on industrial controllers or other software platforms that are trailing the machine. And what you can see here is that software complexity is growing much faster than software development productivity.

So how can this be addressed? For this, I want to bring up the first practical example, industry example that I have here. this video, you can see the robot box. Robot box is a pick-and-place system that is developed by Krones. Krones is an industry leader, a worldwide industry leader, in the bottle-filling industry. So they are-- they're delivering bottle-filling machines, bottle-filling lines. Their R&D is mainly located in Germany. And well, as you can see here, what the robot box does is pick up some containers of bottles and then react flexibly to the demands that their customers have.

And how does Crohn's really address this complexity of a system like the robot box? What they do is they build models. And they built models on the one-hand side, as you can see on the left-hand side of the entire system, so really high-level models to test the functionality, to do virtual commissioning. And then, also they take specific pieces, specific mechatronic components and then develop subsystem models in tools like Simulink and Simscape, where they can really dive deeply into like what are the forces that occur on a gripper system-- like on this robot box-- or what are the maximum torques that you need to have on your motors and gears.

And actually, these system models and the subsystem component models, they are two aspects of the total lifecycle of the equipment, of the manufacturing machine. But as we said before, digital transformation comes from using models over the entire lifecycle. So these models are then connected to requirements, documents, to use cases, to really make sure that what is modeled and what is developed really addresses the needs and the requirements for these machines, so for this equipment.

And then once you have the models and test the models, implementation comes more or less automatically. So what tools like Simulink or MATLAB offer is generation of executable code, like it may be C/C++ or IEC 61131-3 code that runs on industrial controllers, or also executable code in the form of Docker containers that then run on industrial servers or in the cloud. And we will see more about that later on. And then the same models are used for verification and validation. And not only on the physical system itself, but actually much earlier in the design process with the virtual representation with the models of the system and of the software, and this really leads to a shift to the left of the VNV process to identify and to fix errors much earlier in the design process.

But of course, the systematic use of models doesn't end here. So what we see then is that these models are enhanced and updated with data from the field with operation data and then really developed into specific, digital twins for an asset in operation. For instance, for predictive maintenance or for operation optimization. So many of our customers are using models in MATLAB and Simulink for this workflow, which we call model-based design. But they are actually asking for more. So what they really want to see is that they can add additional functionality, specifically additional software functionality, to their machines in [AUDIO OUT] and from the back operation data, measurement data into their design process-- into their development teams.

So in other words, I would say what they want to offer to their customers is the perpetually-upgradable machine. And well, how do we get there? To explain that further, it makes sense to take this flat view of the life cycle and rearrange it a bit and split it into the development phase on the one hand side, and the operations face on the other hand side, and then connect these two with continuous delivery and continuous feedback.

And that ends up in this picture. And I'm sure this picture is familiar to most of you from agile development. This is actually DevOps, a set of tactics that is typically used for software-only systems in IT. But we believe that if you combine DevOps with models, make it model DevOps, then this is a very valid and very beneficial way of developing entire mechatronics systems and not only the software that runs on this.

So let's switch gears a bit and come to the second example that I mentioned. Atlas Copco is a manufacturer of air compressor systems, and they use models in MATLAB and Simulink for designing the functionality on their air compressors. Like, for instance, for automatically adaptization to the specific environment and operation points of the compressor in operation. So they develop a digital twin that's designed based on these models. But then they also connect these models and integrate these models into their manufacturing environment. So connect them to the CNC machines where the individual pieces for the compressors are manufactured, and so they have digital twins produced.

But they also use this-- sorry. They also use these models and integrate them with these configuration tools, that their sales engineers then use with Atlas Copco's customers, to identify the right-sized and rightly-parameter-based air compression. So digital twins is configured. And last, not least, Atlas Copco has more than 100,000 machines in the field where they can continuously get data, operation data, feedback and feed this back into their models that they run on their servers. And then they use the data connected with these models for predictive maintenance, for optimizing maintenance intervals, and for identifying and solving issues in the field as digital twins is maintained. So I would say, in summary, what Atlas Koepka does is they use models and data systematically over the entire lifecycle, has different digital twins of air compressors, and that results, actually, in more flexible air compressors for their customers, higher quality of the systems, and lower maintenance costs.

So for the last part of my presentation, let us step back to the model DevOps image again. And here, I want to dive a bit deeper into the build piece of this process. So what is behind this build piece? So when using MATLAB and Simulink and developing models in systems-- or in environments like MATLAB and Simulink-- you simulate this behavior on your desktop computer. But then, at the end of the day, you need to deploy them somewhere on a production system. So what we offer here is deployment on various different targets, from embedded systems like real-time deployment, to enterprise systems into the cloud.

And I want to dive a bit deeper specifically into code generation for industrial controllers and for PLCs because as we just saw in Professor Jha's talk, PLCs are only present in the manufacturing industry, so they are the-- well, the PCs where typically the software runs, that automates the factory or the equipment. So what does that mean? How does code generation for industrial controllers, specifically for PLCs look like? Well, so first of all, the basis again is models, models in MATLAB and in Simulink.

So these models could be, for instance, controls models like a PID controller or a state machine. This could also be AI functionality, like for predictive maintenance, also like for reinforcement learning when teaching-- automatically teaching a robot that needs to adapt flexibly to its environment as we saw in the robot box example early on. Then that could be optimization algorithms, for instance for path planning. And, of course, it could also be a digital twin, the model of the plant itself-- of the mechatronics system itself-- that runs in parallel, runs in real time, and is connected to the physical plant.