Innovations in modern automobiles take place in the software of the electrical control devices. And functionalities in the vehicle are more and more linked to each other, especially the control systems, which keep a vehicle on track or regulate the distance. Multiple regulation devices intertwine here.

To specify such complex functions in a useful way, it’s not sufficient anymore to describe them on paper and to order different suppliers to realize them. Cross-linked influences and conceptual faults first become obvious when the function is rebuilt and simulated on the computer. This, for example, is done with Matlab Simulink.

At P3, the modeling of functions and their transformation into software via autocoding are hence assured with the methods of Model-in-the-Loop (MIL) und Software-in-the-Loop (SIL).

This way the function and modeling themselves are tested first and then the automatically generated source code is tested in a simulation environment on the PC, long before the actual control device is available. Finally, the last test on the real target hardware follows as well. “By doing this we can find errors even before prototypes of the control device exist – this way errors can be fixed much earlier and the development process can be shortened significantly”, explains software tester Diana Wienand.

The model and the single sub-models have to be checked for their limit value behavior, meta-stable states etc. with different input values. This way we can assure that the output values act correctly in any operational state. And they only do that, if both the single software modules and the interaction of the modules function properly.

“For MIL and SIL we don’t only need technical know-how, e.g. of a gearbox and its functionality, but also deep mathematical knowledge plus logics and analytics”, concludes the graduated techno mathematician.