In this episode of the Future Car Podcast, Nand Kochhar and Gwen Van Vugt discuss the challenges of scaling automotive software and further explore the role of simulations in testing AV and ADAS systems.
Nand Kochhar: There’s a lot of progress over the years as Gwen mentioned in the opening. So, let me Gwen, start with even with these improvements in simulation technologies we are seeing, there’s still an important role of real-world testing to play, right? How do we combine this real-world and virtual testing data, and what benefits can this provide to the overall process of testing and verifying on ADAS and AV systems? You touched on some of it during your previous dialogue, but that’s again one of the key differentiators. Can you expand on that a little bit?
Gwen Van Vugt: So real-world testing is and will remain a very, very important part of testing. I mean, the real-world is extremely stochastic and even though you try to create as many critical scenarios as possible through all the process that we just mentioned, the real world is always a little bit more tricky and more complex than what you can simulate even and what you can think of.
So, in that sense, real-world testing is always important. Another reason why real-world testing is important is also to prove to the public that what you have developed is working in the real-world. People want to see it in the real-world, right? I mean you can show as many simulations as possible but that does not compel as good as just showing that you have driven so many miles without any problems. I think that is all a big part of the validation process.
And don’t forget that in that entire chain there’s also the test track testing and that is also a simplification of the real world, so you take out as much as possible. Yeah, a parameter of the real world in the test track, but what you typically do is you test and also validate your simulation on a on a real-world on the test track. The test track is a very important part to make that link between your simulation models and your real-world testing activities, and therefore it’s not a matter of or, it’s a matter of and.
So yes, you will collect data on the real streets, you will collect data from your simulations, you will collect data from your test track and that all combined together will show you the coverage of your requirements. In the end, you want to prove systematically that all the requirements that you set in the beginning of your product development, that in the end you can meet all those requirements and whether you prove that with real-world testing synthetic data or proving ground testing, you need to be able to cover all the requirements in there, and that’s also part of our tool chain too, to feedback all the answer results to feed them back and connect them to the requirements that you begin with.
Nand Kochhar: So, Gwen, what is required to tie these processes together into a continuous closed loop? And to see this accelerating the verification and validation of AV systems in the future, which is the end goal for our customers.
Gwen Van Vugt: ADAS and AV systems are a very good example of data-driven product development. There’s a lot of data being used in the development of ideas and systems. There’s definitely the concept of the closed loop. So typically, you start with developing your vehicle. You do your test and validation and, as we just explained, there’s real-world data being used. At some point you send your vehicles to the street and then what you see more and more is that these vehicles that go into the street become data generators as well. That can be IoT data, but it can also be very high bandwidth data, and that to collect a lot of data for specific critical scenarios that you’ve encountered in the real world.
What is important to close that loop is the data that you collect from your production vehicles. You must do that in a smart way because even with future technologies you cannot continuously have these vehicles stream all their data. There needs to be intelligence in the car to actually figure out whether the situation that the car is in today is a critical or a novel situation that has actually not been taken into account in the design process.
That, again with the novelty indicator, you could imagine that there’s also technology in the car that continuously verifies, am I now in a situation that is part of my validation activities and therefore I’m in in control and I’m in the area or in the operation design domain that so I’m operating within the operation? Or am I seeing activities that I have not considered before and that could be a trigger to say hey, let’s capture the sensor data of this vehicle and package it and send it out to the back office to take that into account and add that to the scenario database for future developments.
We are all known to the concept of over the air updates, so every vehicle that’s getting onto the street today is set up with over the air update capability. This feedback loop from data from the real-world feeding that back into your design process and your validation process. That’s a topic that manufacturers today are working on.
Nand Kochhar: That’s a very good point. So instead of running prototype vehicles with the special drivers, the production vehicles are equipped with all these technologies of camera writer and sensors to collect the data so you’re leveraging if they’re connected for over the air transfer of data you continuous collect the new scenario, which you probably can’t imagine otherwise and then feeding those back into your database.
I’m sure, Gwen, there’s a lot of machine learning and AI technologies because you’re dealing with tremendous amount of information and parsing which one information is beneficial to update in your database. Now your database is going to be a key challenge and, in these technologies, might be helping in that arena so that’s very important. It’s a continuous improvement from these scenario modeling’s and from delivering functionality feeding towards safety and security of vehicle operations so that’s a really important point.
Conor Peick: Well, as we have seemingly come full circle, I think it’s a good time to wrap up our discussion for today. Thank you, guys, so much for your thoughts and I think we’ve got a good handle on the topic. I want to give you guys a little bit of space to share any final thoughts with the listeners.
Gwen Van Vugt: Thanks for giving me the opportunity. What I think is interesting to share is if I observe the trajectory of this industry. I was talking about the last 15 years and where we’re going, I’m better reminded of the of the common path that emerging technologies of often threat. So, today we have clearly navigated beyond the hype phase. I think the last time that I was in the podcast we were on top of that hype curve, and everybody was really mind-blown about autonomous vehicles. We have passed that, and it is very natural to witness companies evolving, some facing challenges. Also, we have also seen some strategic UM movements into this space, companies stopping companies going bankrupt and also companies taking different paths while others explore alternative avenues.
And it’s interesting because these shifts are intrinsic. I think to the maturity of the of any industry and how we are unmistakably converging towards a maturing industry landscape. You can see that because processes are harmonizing, different standards are getting aligned. One of the best examples is the UN ECE that is picking up more and more of the ADAS technologies into their standards and regulatory frameworks are taking shape and that altogether with the industrial continuous quality cycles that we just discussed at this this closed loop activity of continuously taking data and making your products better. I think that all shows that we are moving towards a mature market and now, in the near future, I think this convergence will lead to the seamless expansion of ADAS across all vehicle types and geographies and coupled also with the growth of higher levels of automation.
And yeah, I think that that’s wholesome. And very interesting problems for society and like preceding technological revolution, such that, such as the advent of, of the Internet, I think these advancements have the potential to reshape the way we live beyond our current imagination. I really think that we are moving towards this industry becoming mature and we will. This technology will be spread out intensively over the next decade and we cannot even think of how that will change our lives and how it will change for example, city planning and all these types of things, it will have a major impact on the way we move and how mobility is set up.
And to be honest with you, I think that it’s great to be part of that type of industry and the work day and night on helping the automotive industry shape our future world basically.
Nand Kochhar: Yeah. So, from my perspective, Connor and Gwen, thanks a lot. Very good discussion. I’ll summarize that you heard about all the challenges and the progress we’ve been making. One important aspect is partnerships. All these things we can’t do are things that companies cannot do on their own. It’s a very important aspect. For example, in our case, we will have partnerships with technology partners within the ADAS AV development phases as well as there’s a heavy compute. I’m in wall and this so we’ll partner with any of the computing vendors, like a WS or other service providers. So, it is important that these challenges are real. We are enabling development as well as verification, validation of delivering autonomy and ADAS and then we partner with for our customers, the partners of their choice to help them deliver this. I just wanted to touch on that aspect as well. Once again, thank you.
Conor Peick: Awesome. Well, I think that’s it for us today. Thank you guys so much again for your time today. Gwen, we really appreciate you sharing your knowledge and expertise on this topic. A quick thank you to our listeners as well who keep tuning in and listening to us chat about the future of mobility and all the promising technologies that that involved. We look forward to having you guys back in the next one. Thanks very much.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.