Proceedings 8th Modelica Conference, Dresden, Germany, March 20-22, 2011 Simulation-based development of automotive control software with Modelica Emmanuel Chrisofakis1, Andreas Junghanns2, Christian Kehrer3, Anton Rink1 1Daimler AG, 70546 Stuttgart 2QTronic GmbH, Alt-Moabit 91a, 10559 Berlin 3ITI GmbH, Webergasse 1, 01067 Dresden {emmanuel.chrisofakis, anton.rink}@daimler.com,
[email protected],
[email protected] Abstract development of control algorithms. This paper presents technology targeted toward the late stages in We present and discuss the Modelica-based develop- the development process, like tuning, validating and ment environment currently used by Daimler to de- debugging the entire controller software in closed velop powertrain control software for passenger cars. loop with simulated plant models. Virtualizing these Besides well calibrated vehicle models, the environ- later engineering tasks requires plant models with in- ment supports automotive standards such as A2L, creasingly higher quality (physical effects modeled MDF, CAN, and XCP to integrate control software and quality of calibration) and near-production con- and simulated vehicles on Windows PCs. troller software (percentage of the controller soft- Keywords: automotive software development, soft- ware included, parameterization using production ware in the loop parameter sets and adaptation of the software to the plant) to be coupled. A tool-chain supporting such coupling should 1 Introduction • be easy to set up and use by automotive de- velopers who are usually not computer sci- More and more automotive functions are implemen- entists ted using software. Hence, there is an increasing de- • support many of the engineering tasks usu- mand to support the corresponding development pro- ally performed with physical prototypes to cess using virtual, i.