Multicore and Timing Analysis

Multicore electronic control units (ECUs) provide additional computational power and the ability for parallel execution, which can be featured by car manufacturers to enable high scale integration of in-vehicle applications, e.g., engine control. Applications, which were initially executed on a single-core, have to be partitioned for integration on multiple cores. Our goal in this project is to provide techniques for migration of control applications. These techniques ensure that approved critical or non-critical single core applications behave in the same way on multiple cores.

In this project, different migration scenarios have been investigated. One of these is the model enrichment scenario, which enables the migration of behavioral models, e.g. models developed in ASCET. The migration is realized by additionally providing requirements specification on the base of a developed architecture specification language. The language is then used to verify behavioral requirements of model-based applications during model-based software integration.

The second part of the project deals with a timing analysis of the models. At first, models are created which contain information of the supplier and the OEM-code. Then, we analyze the timing properties of these models in order to forecast some of the timing problems that may arise due to the parallel execution of the software applications.