This framework enables statistical validation for automotive vehicle simulaitons in large application areas using uncertainty learning (VVUQ_Framework@VVUQFramework_Main). With the framework it is possible to flexibly validate every single simulation result of every individual application parameter configuration. It is focused on complex total vehicle simulations with a large number of changing parameters. The corresponding publicaiton for this framework is [1] and includes detailed information. The aim of this framework is to estimate the reliability of a model and each single simulation result depending on arbitrary parameter configurations of a large application domain.
To get started the newest release of the repository needs to be downloaded including the submodules and the large file storage (LFS) files. It needs to be mentioned that the LFS files are not included when downloading via the zip button on GitHub. They are included when cloning the repository or downloading a release. When opening in Matlab the repository must be added to the Matlab path. To start the a simulation, run the script "Run_VVUQ_Framework.m". The script opens the longitudinal simulation model SmartFortwo, parameterizes it, runs it and shows the results.
To run the code Matlab 2020b or newer is required. Matlab can be downloaded here. Additionally the repository needs to be downloaded at VVUQ-Framework
After the repository is downloaded at VVUQ-Framework it needs to be unpacked in a folder of choice. All Submodules and LFS files must be included.
To run the code Matlab must be started. Browse to the folder where the repository was unpacked and add the folder to the Matlab Path. By running the script "Run_VVUQ_Framework.m" the simulation starts. To run the script, type following command in the command line:
>> Run_VVUQ_Framework
Because the running the Framework takes a long time some evaluation results are already stored in the folder "VVUQ_Framework\Results_VVUQ".
For contributing to the code please contact [email protected].
R1.0 release in line with publication [1].
Benedikt Danquah, Markus Lienkamp, Johannes Rühm, Marius Kirchgeorg, Yasin Meral, Akhmadjon Islambekov.
The Authors want to thank research associates of the Institute of Automotive Engineering, who reviewed and gave advise for the component library.
This project is licensed under the LGPL License - see the LICENSE.md file for details
[1] Benedikt Danquah, Stefan Riedmaier, Yasin Meral, and Markus Lienkamp, “Statistical Validation Framework for Automotive Vehicle Simulations Using Uncertainty Learning”, Applied Sciences, Volume 11, Issue 5, 2021.
[2] Benedikt Danquah, Alexander Koch, Tony Weiß, Markus Lienkamp, Modular, Open Source Simulation Approach: Application to Design and Analyze Electric Vehicles, In 2019 Fourteenth International Conference on Ecological Vehicles and Renewable Energies (EVER), Monaco, 2019.