Defining the Future of Sustainable Mobility

Good morning, everyone. I hope I'm audible at the back. So what I do, I have taken up a couple of trends in terms of how MathWorks in general and such kind of tools in specific have been helpful in terms of the overall automotive journey.

But prior to that, as I understand, this is the first such automotive conference for MathWorks in India, so many congratulations to the MathWorks team to organize this. I'm looking forward to many such-- more conferences in future.

Now, essentially, before I go to the overall flow of my presentation, which is basically capturing at a very high level-- I think you will have a lot of technical sessions during the day. What I wanted to give was the OEM perspective in terms of how we are looking at some of the technology disruptions, which are happening and also taking a cue in terms of how Tata Motors' own journey, especially in the electric vehicles, has been an inspiration for us.

So just to give you a background in terms of Tata Motors, say a journey for the electric vehicles, we started many, many years ago in terms of laying down the vision and also the building blocks for the electric vehicles, and that journey has been very, very fruitful in terms of what you-- probably all of you have seen in terms of our flagship models, both for the passenger vehicles and the commercial vehicles.

So we have truly taken this up to the next level in terms of how you can actually have electric vehicles which are mass-market-ready, both for the cargo and for the passenger segments. So our own journey actually has been an inspiration in terms of how we want to take this ahead in terms of our overall journey and also in terms of what has been our learning.

So at a high level, if you really look at the broad industry trends, this is all about the power of zero's. So when I say "power of zero," it essentially means you would have been hearing it is all about sustainability. It is all about net zero and that has been the defining feature in terms of how as vehicles-- and when we say "vehicles," then the subsystems need to be defined.

So taking a clue from net zero in terms of the power of zeros, I'll just show you a broad heading in terms of what do zeros mean like. So essentially, anything you hear as part of automotive solutions or in part of, say, any aggregate-level solutions, those basically classify in terms of those six zeros.

Now, essentially, if we see the zero accidents, now, a lot of work has been done in terms of how the-- there are no accidents in terms of the collisions. There are good, say, ADAS systems in terms of the active safety. So this is one of the defining features in terms of how, as automotive vision, we can actually ensure zero accidents, and basically, autonomous vehicles and the overall communication, what you see between the vehicles-- that is basically defining in terms of how the vision of zero accidents would actually become a reality.

Zero emissions that you have-- all of you have seen in terms of-- a lot of developments happening around these zero-emission vehicles. So essentially, as we move from ICE vehicles to battery-electric vehicles to fuel cell vehicles, this transformation journey is happening as we speak.

So this is one of the major transformation in the mobility segment, I would say, whereby a lot of changes are happening, and essentially we are seeing scope 1, which is the direct emission, scope 2, which is the indirect, say, emissions, and also, sustainable circular economy. So these are the three trends which essentially would drive the zero emissions.

Zero energy-- now, when we talk of zero emissions or migration to electric vehicles, we also need to see that the energy sources are renewable because if we continue using fossil fuels for the energy sources, then the overall aim of zero emissions does not get solved. So if we have renewable energy, you would actually have net positive, and that's why the word "zero energy." So a lot of effort is happening in terms of how you can actually have a renewable energy driving the zero-emission vehicles in future.

The next is the zero conjestion. This is, again, a collaborative traffic management system in terms of how vehicle to vehicle, vehicle to infrastructure, vehicle to passenger, or vehicle to everything in terms of how you can have a collaborative approach in terms of how vehicles are driven. And then that, essentially, would drive zero congestion. So this is one of the driving factors in terms of how vehicle connectivity and some of the features around that are driving some of the new developments which are happening.

So zero conjestion is something I need in terms of how, whether it is electric vehicles or, in general, the ICE vehicles. This is one of the hallmark need in terms of how this has to be taken a head.

Zero empty-- this essentially means-- today, if you see the way you drive your vehicle, most of the times you are either driving or maybe there is a co-passenger. But it is generally not full, so that is in terms of the passenger vehicle. Similarly for cargo vehicles, if you see, there are constraints and concerns from the operators that they don't run vehicles full-load.

So this is more in terms of how shared mobility would actually drive full consumption of the vehicle capability both for passenger and cargo. So zero empty essentially means the mobility as a service, and shared mobility would drive this zero empty broader vision in terms of what we want to drive.

And zero cost essentially means that the incremental cost of new features would essentially be zero. Say, for example, once the over-the-air updates software-defined vehicles become a reality, the incremental cost of adding a new feature would actually be zero because it could be one too many.

So these are the six, I would say, broader visions in terms of these sustainable disruptions which are driving all our solutions in terms of how we want to look at vehicles of future. Now, essentially, this is at a vehicle level. Now, if you were to translate these requirements in terms of what it means for the computational tools like MathWorks and some of their peers, so broadly it would look like eight major trends.

What we saw in the earlier speaker presentation also, the trend is to move from SIL, MIL, and HIL to driver and loop analysis, which is basically looking at not only the overall vehicle but also how the driver is going to be interacting with the vehicle. So new mode of automotive testing-- basically, the driver and loop analysis is one of the reality in terms of how we are seeing the shift from the way we use the computational tools today.

One of the most important highlight, the way electric vehicles and some of the new technology developments has happened, is the trend about the digital twins. So digital twin basically means you have a digital replica of your physical world, but what it allows-- that while your physical world representation would mostly be static. The digital equivalent actually allows you that you have a lot of dynamic information which can be captured, and on the go, a lot of analysis can be done.

But it also throws the challenge that a lot of additional data gets generated, and how do you make measures in terms of how that digital equivalent can actually be more effective? But this is actually driving a lot of new computational tools because digital twin-- say, for example, if it is the battery health monitoring or some of the driveline-related dynamic performance measurement, digital twin becomes very handy and has become more prevalent as we move to next generation of vehicles.

So this is one definite trend we are seeing in terms of how there would need of much more expanded computational tools. Spatial computing, which goes beyond the AR and VR experience-- I think this is becoming a need in terms of how some of the future complex system would need this.

So we are seeing a lot of trend around the spatial computing, especially on our next-generation vehicles, which will essentially be in the scope of bond electrics. All of you must have been seeing a lot of discussions around metaverse. Now, as we move towards the larger visualization in terms of a broader platform experience, that is where we are finding a lot of takeaways from the metaverse development in terms of how some of the solutions actually not only does the computational job in terms of the model based development, but we actually need to populate a lot of visualizations.

And that's where the analogy in terms of what metaverse has been able to do in terms of the web world-- we see a similar trend or a need happening in the computational world also. Deep learning-based plant models-- I think this is already there in terms of how we use the artificial intelligence and some of the model-based development tools, so the linear and non-linear regression analysis.

So this is anyway there as part of the critical mobility applications. Because of the digital twin, like I just said, the need for big data analytics, and the relevant tools who would be able to utilize that kind of data-- that becomes very, very imperative, so whether it is the large-scale adoption of electric vehicles, or the damaged models, or the arrival, which is basically the remaining useful life of some of the aggregates which is a distinctive feature which is different from our ICE vehicles where we never had this kind of a remaining useful life kind of analysis being part of the study.

Quantum computing-- this is, again, with the specifics of new-age vehicles, which is becoming important. Now, in terms of the way you would do a range estimation for electric vehicle-- this is on-the-go, very, very complex regression analysis, unlike our conventional ICE vehicles, because the way a driver drives, the way you have the profiling of the road and the external environment conditions-- all that plays a very, very important role in terms of how dynamically your range estimation happens for, say, an electric vehicle.

That's why the traditional tools will not be sufficient in terms of the way some of the dynamic measurements are possible or are currently being done. So quantum computing-based-- some of the solutions is something we are looking at, and this is both at the hardware and at the software level.

And lastly, what we have already seen in terms of some of the tools being able to provide that, which is the mixed-reality applications, like how closely you can represent your physical sensors-- so broadly, this is what we see as a broader trend in terms of the computational tools, and, essentially, MathWorks is there in terms of most of the solutions. And we have ongoing discussion in terms of how some of the things become relevant as we see the trend or the need for future vehicles.

Now, having seen the overall vehicle level, say, advancement you have seen in terms of the power of zeros, then what it means in terms of the overall need for computational tools-- now, if we see, both these trends actually need a lot of competency mapping to address these trends. The conventional way of working in terms of whatever knowledge base we had-- so this is, again, our OEM perspective, I would say, that we are finding a lot of needs, and the earlier speaker also shared some thoughts in terms of how we can actually bridge the competency gaps.

So essentially, if you see, basis electrification, basis connectivity, basis-- the ADAS and the overall autonomous vehicles, some of the R&D challenges are basically in terms of the real-time data processing in terms of the safety consideration and the complex integration and some of the new thermodynamics in terms of especially the fuel cell in electric vehicles.

So there are skills required which are there in terms of what people have, but in terms of how it goes to the next level so that the upcoming solutions can be better-served-- so there is sort of a summary in terms of deep learning model-based testing and overall training. NLP, the natural language processing, is something which we are seeing a lot of trend in terms of extracting the data and being able to analyze it and overall end-to-end analysis.

So this is in terms of the skills. Now, as you upgrade your skills, you also need improvement or, say, extension of the infrastructure, which is basically in terms of the new tools for modulation and simulation. The charging infrastructure actually, along with the vehicle, poses completely different challenge. So you have thousands of charging infrastructure available in country today.

And the intent is all those would be connected real-time. So as you drive, you would get a real-time emphasis or the feedback not from a perspective where it is available but whether it is available in terms of the right capacity also.

So Realtime Database, which was not the need for ICE vehicles-- that's something relevant for today. And also in terms of the testing infrastructure for the balance of plant, which is very specific to the fuel cell vehicles. So broadly, if you see, the digital product creation-- we have measured the overall flow in terms of how the requirements can be met and how the virtual flow basically happens. So that's a small sort of a snippet in terms of the digital product creation.

And lastly, I'll show in terms of a very small example of how we have been able to use MathWorks toolchain. So in terms of the tools which traditionally we have been using-- those are the Simscape, Simulink, Polyspace, and some of the embedded coding. So essentially, these tools have been defined for building plant-level models for the base software, for the application software, and some of the code verification applications.

And the development flow basically has been in terms of how we can use the embedded code for the code generation, so the Simulink Test for the unit and overall integration testing, and basically using, traditionally, the MIL and the HIL. We basically see a lot of need in terms of the driver and loop analysis going forward and, essentially, the damage models or the remaining useful life models for the electric vehicle aggregates going forward.

Some of the typical use cases have been traceability of requirements. We have been extensively using model-based system engineering as part of the overall use cases, some of the traceability of code to model, and, essentially, agile way of development, and also the functional safety. These have been the two new models which we have been looking at.

And we are also looking at some of the larger improvements or, say, work in progress in terms of how you can actually do a mode optimisation both for the real asset of the ECU in terms of where model-based developments sort of integrate and also in terms of the asset, which is actually required for compilation.

How do we get better integration for some third-party tools like Git-it and all? How do we make the outboard profiling more accurate? And broadly, the integration of next-generation functional safety models so that the overall, say, compliance to ISO 26262 is better-managed, including some of the aspects of cybersecurity--

So broadly, this has been our broad journey, but there are a lot of, say, differentiation happening to basically cater to the next generation of vehicles, like I described in my earlier discussion.

So this was broadly at a high level I wanted to show in terms of how we are seeing some of the vehicle trends, what those vehicle trends mean in terms of the need for computational tools, what sort of gap we see in terms of the capability and the required infrastructure, and a very small snippet in terms of how, as Tata Motors, we have been traditionally using MathWorks for our generic function.

Then, like I started, we draw a lot of inspiration from Tata Motors' own journey of electric vehicles and look forward to long associations with MathWorks in terms of how the traditional tools and what we see as a future will have a good handshake. So thank you, and looking forward to good sessions, and all the best to all of you. Thank you.