How Nissan is Accelerating its Vehicle Development Cycle with Sonatus

0:00 Introduction to the Podcast

Welcome to another episode of the Driving Innovation Podcast.

Modern vehicles are amazing feats of design and engineering, but they’re also increasingly complex to develop, test and validate.

Today, we’re diving into how one of the world’s leading automakers is revamping its pre-production validation process through a data driven and AI-aided approach to simplify that complexity, gain efficiency and ultimately save costs.

I’m your host, Sanjay Khatri, and I’m thrilled to introduce our guest, Sarah Dorey, Senior Engineer for e-Planning, Coding and Cybersecurity at the Nissan Technical Center Europe. Sarah is here to discuss how the center is solving major validation challenges and accelerating its vehicle development programs with the help of Sonatus Collector AI and AI Technician.

We’ll explore the power of event driven data collection, the insights unlocked by connecting data across complex vehicle subsystems, and how an iterative closed loop workflow is fundamentally changing how they conduct root cause analysis.

Sarah, thanks for joining me on this podcast.

Hi, great to meet you. It’s lovely to be here again.

1:24 Validation Challenges in Vehicle Development

Thank you. Yeah, so let’s get right into it. Your organization is at the center of Nissan’s vehicle development program for, I should say, award winning and popular models like the Nissan LEAF and Juke. By the way, congratulations on Nissan LEAF getting all sorts of awards. What were the validation challenges your teams were running into that made you look for a different approach?

Yeah, thank you. Great question, because it’s something that’s been challenging us and many OEMs through the development phase.

We tend to find difficulties when it comes to root cause investigation in all of our standard developments.

It’s nothing unusual within the industry, but with the distances where our manufacturing plant is at two R&D centers in Europe, one in the UK and one in Spain, gathering data can be really difficult. You have to be on the spot or at least have somebody with the skills, knowledge and capability in the right place at the right time. And that’s not always the case. Quite often you’re gathering data, you go back, you analyze the data. And even after that analysis, you want to do some more data collection to really dig deep into the data, really analyze what ECU is having what problem with what scenario or what CAN connection. So more data logs are required, which if you’ve already left the site is a bit of a problem.

Also, it comes to our development cycle and we’re adding vehicles go down the production line, we have problems come up as they’re produced.

The traditional model is that the data loggers go on as the twelve volt battery is connected.

But we don’t find out what’s on that data logger until it reaches the end of the line, somebody collects it, they upload the data and then notifies us that it’s available in the cloud so that we can go into that data again.

So this kind of slows down the investigation process and makes it quite lengthy. We want to speed up with what we’re doing. We have such great products. We want to be able to get them out to our customers a lot quicker.

The products are a lot more complex as well. So there’s a lot of development concerns that we need to have a look at, which are standard concerns. But we want to make sure that the customer doesn’t get to see these.

The software is so complex. You’ve got cybersecurity on top of that. There’s a lot to dig into and we want to make sure that by the time we release that product, that the customer is getting the best of what our products can deliver. So it’s really been able to speed it up and reinforce the quality that we offer our customers.

4:10 Digital Transformation in Validation Processes

Yeah, so I hear you describing something which maybe I can characterize it as digitalizing or virtualizing the process so that it’s more dynamic, it’s more real time without having to wait on assets, wait on test vehicles, do a lot of sort of manual transfers and so forth. So I mean, it’s reducing a lot of the friction in that process. So you’ve adopted a more sort of digitalized and AI enabled workflow with Collector AI and AI Technician. What sorts of results have you seen?

Yeah, it’s really quite promising. We’re obviously going through our proof of concept project at the moment, but the engineers are really enthusiastic about the opportunities. As I said, traditionally we’re having to travel to go and collect data.

With the remote deployment of policies, we don’t need to be able to do that. And then with the redeployment of policies, again, it’s very quick and very easy.

We can see results really quickly. When we then turn around that data into AI Technician, then we can start to go through the FTA (fault tree analysis) really quickly. We can start using our experience and our education to be able to ask the questions that go into the data that AI Technician connects all the dots for us. So we’re not having to sift through different SharePoints and different files to pull all that knowledge together or go and speak to people across the business because we’ve pooled all of our Nissan knowledge and using AI Technician, we can bring it all together.

So it’s really exciting that the speed that we can be able to do this. And we need all of our engineers still within our business to be able to make these decisions, but they can at least do it quicker with more access to data and real reliable data as well. So it’s a real bonus for us.

6:04 Remote Access and Global Collaboration

So I hear you saying, number one, it allows your engineers to work on vehicles without having physical access to vehicles. So you sort of remove that dependency of contention with resources in terms of your test vehicles. And then the other one is being able to sort of dynamically do things potentially even in a test drive, I would assume, so that you’re not having to kind of bring the vehicle back, refit it, adjust the loggers that are in the vehicle and so forth. So is that how you would characterize it? And how is that helping sort of, you talked about kind of a global team with multiple sites, presuming you have interactions with folks in Japan as well. How is that all helping you collectively?

Yeah, it gives us a lot of flexibility and agility to react to situations.

This ability to remotely deploy the policies means that we can have a car going around to test track in Barcelona, but I can be accessing it here my home. I can be doing it from the office. Somebody from Japan could be checking the dashboard data as well. So this one car and one data set with access for all really opens up the opportunities for us to go quite wide across our test fleet when we’re testing in a way that we haven’t necessarily done before. Normally, test cars are allocated to a particular function and a particular test. So it really does mean that we have more resources for all of the engineers to be able to use across the globe if we’re adding this remote access. So that’s really quite exciting.

Let’s talk about the analysis.

As I mentioned early on, vehicles are amazing these days, including the ones that I mentioned. But they’re also very complex, a lot of electronics, a lot of what I would call sort of cross domain dependencies, features span multiple subsystems.

And oftentimes what I would assume is that connecting a lot of those dots across those subsystems can be challenging, especially when you’re looking at one particular data silo.

How has AI Technician, in conjunction with Collector AI, helped your engineers connect the dots across these domains? And what sort of insights and time savings has that unlocked?

8:33 AI Technician’s Role in Data Analysis

The engineers, their skills are exceptional, but they can’t be an expert in absolutely every ECU and every system across such complex vehicles. And as I said before, once you add cybersecurity on top of that, then it really does add a new level to what we’re working on.

To bring in AI Technician that links with the Nissan Knowledge Data Lake, AI Technician can bring that knowledge together. It can make everybody an expert.

You don’t have to be waiting for an email from somebody for a few days. We’ve essentially gone for if we’re root cause investigation on a particular technical concern, we can see where the opportunity is. During our proof of concept delivery, we’ve seen that we can go from two weeks to two days for a root cause investigation. And that’s massive.

But because it breaks through that complexity, because it can join all of those dots for our engineers, that’s really exciting. I absolutely love how it comes back with my Nissan knowledge because I’m an engineer. I don’t trust anybody. If you’re going to give me an answer, I want proof.

And you’ve got to go really deep to make me believe it. And the Technician Solution does It comes back to me in my own language. It starts quoting documents at me that I probably wrote at some point and it goes into all of that depth that actually says, this is why you should believe me. But it also tells me why it discounted other things. That feels like a conversation that I’m having with our engineers. And that therefore makes me believe that it’s really gone into the depth of the knowledge that we have at Nissan and we have many years to be able to call on to give me a good quality response back so that engineers go into that decision making process with a real rich knowledge backing up their decision making, which is where we want to be. That’s what we do. But now it’s time consuming. With the Sonatus AI we can get that through that so much quicker.

Interesting. Yeah, I like how it’s it’s augmenting the work that engineers are doing by giving them almost an assistant on the side who can go out and get that information that they may not have access to, or they’re not sort of inherently knowledgeable about. It sounds to me also that there’s some sort of of an iterative process, right? So rather than sort of a static linear, you get some data, you do some analysis, you go back and test, and it’s sort of a sequential process.

11:10 Closed Loop Workflow Dynamics

Can you describe to me some of the sort of the closed loop workflow aspects of how sort of this dynamic data, the analysis of AI Technician, and then potentially kind of going back and iterating on it? Can you give me a little bit of insight in how prevalent something like that is and how much of that is actually helping you?

Yeah, and engineers are the kids that always grew up saying, “Why”. Why, why does this happen? So whenever a problem comes up, a concern needs investigating, that’s what they’re going to do. You give me some data, but then I’ve got more whys to go into. We set our policies on that.

It can be quite broad, so we can just ask it to look for any DTCs that come up on an ECU. And it will flag that and it will collect that data automatically. That’s really cool. But what we can also do is pick up CAN signals that we specifically want to target.

So if I know I’m looking for something related to that is on the CAN network, then I can set that and it can bring that back to me as well. So I can have these simultaneous data collections to really cover everything that I’m looking to do. It’ll upload all of those data logs into AWS for me. So then it sat up there into the AWS with our Nissan knowledge and they sit there side by side.

And with AI Technician, we can then start interrogating that. The cool bit after that is that we can take that FTA knowledge, all of that “why”, and put it back into a policy and redeploy. So from our user interface that here in my garden or the guys in Spain or Japan globally, we can just redeploy that. And we can keep going on this loop.

With the closed loop system, then we’ve got that additional safety. And Nissan knowledge isn’t sat in somebody else’s cloud. So there’s no vulnerability when it comes to that element. So we can feel quite secure that all of our knowledge is held together, whilst being able to really deep dive into our knowledge.

So it’s really cool, really.

13:19 Impact on Cost and Efficiency

You talked a little bit about those better alignment across global teams.

As you expand these tools to programs like the upcoming Juke and the LEAF Models, what do you see as the next step in the smart testing AI assisted validation?

Yeah, thank you. It’s a real opportunity now to take more of our digitalization work and feed that back into this whole process. And in the proof of concept, we can see where the advantages are, and off we go. We can then get those recommendations from the [AI] Technician that give us other opportunities, whether it’s seeing trends across the business where our fleet is flagging certain anomalies. Or if we’re just looking to secure that data across our whole fleet, can do one vehicle, we can do many vehicles. So we can then begin to look where we can do simultaneous testing.

So it might be me needing a data set. It might be two or three of my colleagues that also need a data set from the same car. We can set our policies onto that one vehicle. We don’t need three different vehicles.

So that’s where we’ve been able to look at that reduction in the test vehicles. The cost is high, they’re prototype vehicles. So any advantage for us to be able to do that is always going to be a win.

Well, has been really helpful, Sarah. And I really want to thank you for sharing your precious time. I know you’re very busy, but this has really been helpful. And I’ve learned a lot. I thought I knew a lot about project that we have going on, but I’ve also certainly learned a lot and you’ve added a lot more color to it. So I really want to thank you for joining us.

Thank you. Thanks for having me. It’s always great to chat with you guys.

15:16 Conclusion and Key Takeaways

That brings us to the end of another insightful episode of the Driving Innovation Podcast. We’ve seen how Nissan Technical Center Europe with the help of Sonatus Collector AI and AI Technician is fundamentally changing pre-production testing and validation. By leveraging event driven data collection and closed loop AI assisted workflows, Nissan Europe is accelerating vehicle development, reducing costs, and achieving better global team alignment. A huge thanks to Sarah Dorey for sharing how this smart testing approach is shaping the future of models like the Juke and the LEAF, And we look forward to bringing you more of what’s next in the evolution of AI assisted vehicle development and validation in future episodes.