Last week I was fortunate to be able to attend the Automobil Elektronik Kongress in Ludwigsburg, Germany. The capabilities of the latest luxury vehicles presented by BMW, Audi and Daimler were impressive to say the least. You can read a little about the upcoming Audi A8 in this review here but the Mercedes-Benz S-Class will have all the bells and whistles including a stereo camera, detection of pedestrians, rear collision monitoring, road surface scanning, lane departure warnings, blind spot detectors … the list is endless and impressive.
As you may have noted from the models I mentioned all this advanced driver safety is making it’s way to us through the top end of luxury lines. And it’s no wonder. According to the presentation given last week by Dr Dannenberg of Berylls Strategy Advisors in 2025 the value of E/E systems in Luxury vehicles is expected to be $40,000 while the value for Premium vehicles is expected to be over $15,000 and for Volume production vehicles a much lower amount under $5,000 per vehicle.
What this does mean is for non-luxury and non-premium models these advances will likely remain options for a while. But all hope isn’t lost. A number of things are making the development and production more cost effective over time. For one thing the use of AUTOSAR makes repartitioning software functions between ECUs or consolidating into fewer ECUs more feasible, and improves the ability to reuse some of the software investment made once in future models. In addition, SoCs are bringing more processing capacity and software is enabling us to separate domains so those that require greater safety can exist on but still be isolated from other sub-system that have different safety requirements.
Harald Kroger, Vice President Electrics/ Electronics & e-Drive at Mercedes-Benz Cars. discussed some of these trends and challenges in his presentation last week as well. He expected to see 15 Million AUTOSAR ECUs by 2014 and 30 Million by 2016. Now the need and move towards consolidation has been discussed for a while and it isn’t simple. Mr. Kroger was asked if he expected to see the ECU count increase or decrease. He had a very appropriate answer that the answer isn’t simple. There is a balance that must be found between the topology and cost. It would be nice to be able to have lots of functions in one ECU but you have to balance the size of the ECU and if it will fit and wiring (cost and layout).
I was discussing a simple example of this with a colleague recently, the use case was if it would make sense to consolidate the controllers for two door windows into the center or leave them separate. When modelling this scenario with Mentors VSX tools he found that the added cost of wiring from the center out to the two windows made it more economical to have separate controllers. That seems logical enough. Of course a modern car isn’t just two simple controllers. With close to 100 ECUs in some vehicles it is a cross discipline design that requires connecting the electrical, mechanical, thermal and embedded software domains. And this connection is what we’re continuing to work on in our teams, it’s good to see there some tough problems out there to keep us all busy