The Complexities of modern automotive software – The Future Car Podcast – Transcript
In this blog you can read a transcript of our recent podcast on the complexities of automotive software development today and in the future. You listen to the podcast here!
Conor Peick: Software is now a critical component to nearly every modern vehicle. Consumers have demanded more and more high-tech features being incorporated into their vehicles, and software content has grown as a result.
This transition towards a software defined vehicle has also caused a rapid growth in vehicle complexity to keep up legacy automotive original equipment manufacturers and suppliers must be able to deliver innovative software driven vehicle features while integrating these features into the mechanical, electrical and electronic systems of a vehicle.
Today on the Future car podcast, NAND and I are joined by Piyush Karkare, Global director of Automotive Industry Solutions at Siemens Digital Industries Software. We’ve invited Piyush on to help us dive into the complexities of automotive software development today and in the future, particularly as we come closer to the reality of autonomous vehicles.
So welcome, Piyush. It’s great to have you on and you know if you could start out by just giving us a quick summary of your background and your experience in the field of automotive software.
Piyush Karkare: Hey, thanks. Thanks for getting me on the podcast. Yeah. So for my background perspective, I mainly have been in the auto industry for last, I think 29-30 some years now. 17 years of that has been at Ford Motor Company mainly in the electrical and the in vehicle software area. Before I joined in at Siemens about 6 to 6 and 1/2 some years ago. So mainly have been in the electrical and software area for automotive. That has been my background.
Conor Peick: That’s awesome. So you know, Piyush, we’ve talked a lot about the broad trends in the automotive industry. Yeah, these are really affecting the ways in which vehicles are designed and developed. So I’m curious, you know how these trends affected automotive software development in particular and you know NAND, you’re welcome to jump in here as well. I’d like to get pieces perspective first I think.
Piyush Karkare: Yeah, sure, sure. Yeah, automakers now are applying software to obviously meet all the market demands there that we see today. They have found that the software enables them to make these vehicles much smarter, more adaptive and more convenient, even more desirable and allows these allows automakers to accelerate there. You know, development cycles. But obviously software, as you add more software to you’re vehicles. that’s a lot of complexity in particularly the way it has to interact with the hardware that software goes on to.
Nand Kochhar: That’s another key point. Many of the features consumers want today, they would not be possible without software. You know, for example, some of the feature functions software helped companies implement them better, faster, more intelligently, and what have you. Consider the difference between early cruise control, and current adaptive cruise control systems?
And for other feature functions, the software is the enabling piece. The obvious example here is advanced driver assist systems and self-driving systems where software receives the sensor information and triggers responses in a matter of milliseconds. The result of all this is that the software content of vehicles is growing rapidly over 15% compounded annual growth rate, by many estimates.
So, Piyush, I like to get your perspective on what kind of challenges this growth of vehicle software content introduces to the automotive development life cycle today.
Piyush Karkare: Sure, if not anything, you know it’s definitely complexity as you mentioned before, as you integrate more software into the vehicles, obviously it increases enormous amount of complexity that’s coming into the way calls. This is particularly true when you know software and hardware systems have to interact you know think about the software that you want to you know put on your way call, but you have to understand as to is this is the hardware enough to put that software you know, lot of software has to work most of the software. I would say that that goes on the vehicle is real time software. So all the hardware that is on your vehicle has to be able to compensate for that real time aspect of the software and not only that in in terms of the computational power or performance that is needed that if certain thing has to happen in you know a certain 200 milliseconds the hardware has to be compatible for that kind of stuff.
It adds a lot of hardware as well as software complexity and also this added complexity affects the overall vehicle development process as well. In particular, it sort of highlights the siloed nature of development. You know the more complex the system gets, it leads to more frequent and severe problems when engineering teams and disciplines synchronize their designs, it vastly increases the challenge of managing the vehicle development on top of, including, you know, the internal teams and departments and even suppliers, there are tasks with deliver components and subsystems. And because of the siloed nature, it potentially hides a lot of issues and problems only to find out later that you have some that you have some problems in inbox to fix.
You know, as you go more towards the launch and thus you know in all to see it increases the overall vehicle development risk and leads to more recalls that we see nowadays are kind of because of software and in the warranty claims and the cost to the automakers are kind of coming in because of that.
Conor Peick: So when talking about an AV and the software demands such a vehicle will have, how does this change the kind of the software equation for automakers?
Piyush Karkare: Well, AV obviously just kind of extrapolates the complexity right when you’re thinking about, you know, making these machines do what the humans can do, right? So obviously it increases the complexity. All systems are becoming, you know way more complex and sophisticated including all the hardware that needs to go in there, the software, even the mechanical and electromechanical systems, the way they have to interact the software specifically.
You know the complexity is solved in several areas, including what or not limited to design and development, where the software systems will have to fulfill a greater number of requirements for feature functions that are going to go in there, verification validation is a big thing, right? Compliance of the software or you know, all the way from, you know, safety critical systems to, you know meetings, strict requirements and regulatory things for, you know, functional, safety and cybersecurity and things like that change. Management is a big deal when these changes happen, how do you test and how do you validate these changes and cascade them to all the affected systems, remember the software is not, you know, working in the side of software is working on a hardware. The hardware is connected overall with electrical systems. Electrical systems are driving some of the mechanical components and things like that.
So the cars are already out there in the world. How do you implement these changes in the software on those vehicles, you know you need to know is what hardware is on it, how it is connected, and all of that kind of stuff. So it’s changing, you know, from that perspective enormous amount of change happening because of airings.
Nand Kochhar: Yeah, that’s a very good point. I think the automotive business is complex to begin with, but I think overall autonomous vehicle will take the complexity which we see today and just accelerate its growth to a point that’s going to be unsustainable given current development and systems engineering methods. So bottom line, automakers are going to need to look at a new way of doing things to keep on this path today, teams are not equipped to handle these challenges, especially as we move into the increasingly digital future. So what do you think are the biggest holes in current automotive software development methods?
Piyush Karkare: Yeah, very interesting question. In a lot of companies now are adopting what software used to do, which is more of a feature-based process, feature-based approach to overall you know to control this complexity. There are many holes to this whole approach or to the way they are getting on to this approach. And even the companies who are not yet on this feature-based kind of approach.
Some of the challenges that I see, you know, things like scalability, right? Many companies today rely on very document-based approaches. You know this will not meet the, you know, the challenge, the scale will not meet the challenges of the modern car and certainly not an AV, right. Compliance for AV is a big thing as you as you mentioned earlier many companies have to ensure that their software development process is compliant to the functional safety and cyber security regulations.
So how do they do this? You know, while the complexity is growing, change, I mean in our industry and change is a constant thing, right? They have to manage engineering changes as these issues arise. You know these issues could be very much complex and you have to do that very efficiently. Collaboration is a big deal, especially how you share and work with suppliers as many, many, many suppliers are involved in developing and deploying and testing, not only just the software piece of it, but also the hardware side of it. And companies must do this in a very secure and efficient way.
Another one is reuse. You know, how do you ensure that you know you can reuse the functionality that is available on vehicle programs and deploy that to multiple programs. How do you maximize systemically the reuse aspect of these, you know software features or features in general?
Lastly openness. You know, how do you integrate many data, you know many development and testing tools into a robust process. How do you do that? I mean, these are some of the challenges that come to mind.
Conor Peick: So what can automotive companies do to kind of fill in these holes is, you know, is that something that digitalization can help them do? Or, to put it another way, you know, how does digitalization help these automotive teams prepare for the future?
You know again, especially as we get into autonomous vehicles and the complex software systems that that they will require. So, Nand, what do you think about that?
Nand Kochhar: Yeah. So, let’s start at the top level. I think one of the most important things digitalization can do for software is improve the integration and collaboration between various engineering domains, some call it cross-domain so software engineers can work more closely and efficiently with their counterparts in electronics hardware, electrical-electronic systems integration, and the mechanical domains as well. This increased collaboration can reduce the program risk by making it easier and quicker to find and resolve design issues, or by avoiding these problems in the 1st place.
Finally, this can help all domains further improve and optimize their designs more, even on increasingly tight time constraints, as less time is spent on resolving issues during the integration of the various domains. So in other words, the continuous integration during the development process.
Another important thing is that digitalization enables automakers to take a true software and systems engineering approach (we call it SSE) to the overall vehicle development process. A digitalized SSE process means that you can maintain a holistic system-level view of vehicle during its development through connected digital tools. The systems-level view includes various domains and subsystems, including software, hardware, mechanical systems, and all the others we talked about before. Some in the industry refer to this also as a model-based systems engineering as part of it, but we are emphasizing on software. So the SSE will capture product requirements, performance targets, constraints and other information to form the product definition.
These definitions and requirements are then available for all other stakeholders, which includes traditional CAD, CAE, or simulation, and even partner companies such as software suppliers. So you’re engaging the entire ecosystem, bringing under the same level SSE is not a one way flow of information, as simulation and tests are completed, that data can be captured to further enrich the digital twin of the vehicle. So in other words. It allows you the closed loop integration process by feeding that information back and making everything better at every step of the way.
Conor Peick: And so then, Piyush, how do you see this? You know what Nand just described this SSE process? How can you apply that specifically to vehicle software development?
Piyush Karkare: Sure, sure. Yeah, that’s you know, we feel that that is a very you know in demand right now kind of thing where we talk about software and systems engineering is you know it’s one of the biggest things that digitalization can do to help coordinate and orchestrate the overall software development flow from high-level program engineering. You know add feature function, decomposition level doing the create studies to ensure that you know things that you’re conceptually designing. Are they even feasible. Setting the targets and so on and on right do core controls in software engineering processes.
A digitalized software development flow based on SSE, like what we offer from Siemens, typically we talk about in the kind of three pillars. One is requirements and test specifications as to how we assign attributes with an integrated traceability underneath. With an online platform that manages all the workflows underneath and test case management with requirements traceability to it.
Second is design and testing, linking requirements to models and the code that is being written or being degraded. Tracking the best results, and automating that testing aspect of it, and then planning for releases and status monitoring as to where the project is at.
And lastly, reporting and dashboards is providing insights into the overall model development that’s happening and automate those results to feedback back into the requirements and test specifications and design.
So these pillars, if you want to call those you know, create a solution that can continuously provide the right information to the right people at the right time in all phases of development, you know, during the specification implementation and testing phase, while looking for models to reuse and doing the change and impact analysis, the viewing and managing this model information from a global platform is paramount and that helps in the overall, you know, traceability of the relevant models.
And who owns it? You know who’s managing it? You know which people are accessing it. How are they getting reused by all involved stakeholders? All of that is then possible using this kind of approach.
Conor Peick: So, for companies that are, you know, maybe looking into how they can really digitalize their software development, how do they get started, what’s their first step?
Piyush Karkare: Right. How do I get started, right. I mean honestly the reality is that you know migrating from you know sort of the legacy approach that can be very difficult you know to efficiently achieve this end-state of a digitalized systems approach to vehicle software development companies really need to seek a technology partner who really understands the engineering for the controls. So mechatronics and performance engineering and systems engineering.
They need technology on how to effectively use simulation and data management technologies to support processes and engineering activities, and by doing so, the customers can achieve the value that is, that is the seek faster and with the reduced risk during the transition.
Conor Peick: Nand, I want to toss it over to you as well. Do you have any thoughts about how these companies or engineering teams can get started on this on this transformation path?
Nand Kochhar: Yeah. I think digitalization or digital transformation journey, you have to start in an area where you are challenged the most within your business processes and you have to win small success in that area and then grow and expand into the other parts of the business to deliver end to end solution. So I think there is no cookie cutter or A1 answer. It depends on customer situations, what problems they’re facing. But you have to get started on that journey. So what we’ve been calling digitalization or I call it, a digital transformation journey. Once you get started, you can truly convert the software complexity, which was the subject of today, into a competitive advantage for delivering value to your customers rather than in terms of cost savings, whether it’s reusing the product development time or, the most important to me, is the innovation that you can develop next generation of your products.
Conor Peick: Thank you guys so much for joining us today. You know Nand and Piyush, I really appreciate your time and expertise on this topic and you know, and I’d also like to thank our listeners for tuning in and hearing what we have to say about this really fascinating topic. We will be back with more discussions on the future of the automotive and transportation industry, autonomous vehicles and sort of everything that’s involved in that. So thanks again for listening and Nanda, Piyush, great talking to you.
Siemens Xcelerator, the comprehensive and integrated portfolio of software and services from Siemens Digital Industries Software, helps companies of all sizes create and leverage a comprehensive digital twin that provides organizations with new insights, opportunities and levels of automation to drive innovation.
For more information on Siemens Digital Industries Software products and services, visit siemens.com/software or follow us on LinkedIn, Twitter, Facebook and Instagram. Siemens Digital Industries Software – Where today meets tomorrow.
