Creating a drivable car in virtual space



Jonn Lantz
03/02/2015

Automotive IQ recently spoke with Jonn Lantz, Technical Expert Electric Propulsion Systems / CAE, engaged at Volvo Car Group, about software models in agile mechatronics and the evolution of model-based automotive engineering.

Mr. Lantz, In you current role with Volvo what are your main responsibilities?

I work as a technical expert for Volvo. Our department is developing devices in the electrical powertrain like batteries, electric motors, chargers. My key role is to help with development all around to get more agile development. Earlier, I have worked more in the direction of change driver but today I work with continuous integration in our main projects. The modelling part of this work has also grown, now including pre-studies, physical modelling, virtual verification, complete vehicle models…

What is your background?

Actually I am a quantum physicist. I used to work a lot with simulation programs and software. And this works quite well with the automotive industry with all its hardware and software. I also did a lot of modelling. For around seven years I’ve worked for Volvo in the field of introducing model-driven development in various groups.

You will present on Models in Agile Mechatronics at our Embedded Systems conference in Berlin. What are the main challenges of modelling?

We approach this area from a grassroots perspective. First of all, modelling is just an abstraction of c-code. Agile development with models is tricky in a sense that there is no modelling community. There is a C++ community and a Python community but unfortunately there is no modelling community. We need to create one in order to help each other.

Another challenge is that in automotive we really have to test new software on the product. But, since the product usually doesn’t exist; we have to model it first. So we need to have a virtual car to test new software. To create a large, executable model made of models is difficult as you have to integrate models of different kinds. It has to be executable. And the model types can be very different.

We have to create a drivable car in virtual space so that we can replace control software in the model car with new versions and then test. This is our way to approach incremental development.

You focus mainly on electric propulsion systems. What impact did the increasing number of electronics in the car have on electric cars in comparison with regular cars or hybrids?

I worked with both of them but the difference is not so big. In general, hybrid cars introduce more novel electronics. Electric cars are simpler. It’s the hybrid cars that are most complex, as they have two drivetrains in the same car. And these drivetrains have to collaborate. We need to deal here with a lot of new features. New ECUs have to be developed from scratch . We don’t have much legacy here.

What are the main aspects that should be kept in mind in terms of product integration?

Numerous and well defined layers of testing. The modelling comes early in the software development. After the models are converted to c-code the integration will be more or less standard software integration. But you could also approach it from the virtual verification point of view. We try to create a parallel track of virtual integration, with corresponding test levels.

What do you wish and what do you expect for the evolution of modelling/model-based engineering for embedded systems in the future?

I wish for better languages, more flexible modelling languages, more integrated with architecture. But I really expect some development on the integration of models. Let’s say if you want to simulate a system within a system. The dream is to have the virtual car in the cloud. Let’s just download a copy of the current product and then integrate our little software changes within it, run all the legacy test from the cloud and just see whether a new development feature fits or if it fails. A more straightforward integration into larger systems based on cloud services, that’s what I hope to see in the future.

Thank you for your time.

[inlinead]

Impressum :
Firmeninformationen entsprechend § 5 Telemediengesetz
IQPC Gesellschaft für Management Konferenzen mbH
Address: Friedrichstrasse 94, 10117 Berlin
Geschäftsführung: Silke Klaudat, Richard Worden, Michael R. Worden
Telefonnummer: 030 20913 -274
Fax: 49 (0) 30 20 913 240
Email Adresse: info@iqpc.de
Registereintragungen: Amtsgericht Charlottenburg HRB 76720
Umsatzsteuer- Indentifikationsnummer DE210454451