Präsentation am DO, 10.10.2019, 10:30 Uhr, Besprechungsraum ITI
MA-Endpräsentation: Vimal Sivashanmugam B. Eng
Betreuer: Univ.-Prof. Dr.rer.nat. Marcel Baunach, ITI, Graz University of Technology /Dipl.-Ing. Tobias Scheipel, ITI, Graz University of Technology/ Mr. Bhargav Adabala, AVL List GmbH
“Standardization of Control Unit Software with AUTOSAR Norms“
AUTomotive Open Standard ARchitecture (AUTOSAR) consortium has been set-up to promulgate a common standard in software development across the automotive industry. AUTOSAR proposes a unique layered architecture for automotive software and a unique software development methodology. In recent days, the automotive companies have been increasingly adopting AUTOSAR principles for their native software development. While AUTOSAR methodology generally involves the development of software from scratch, the question of whether an existing non-AUTOSAR software can be converted to AUTOSAR format has also been of greater interest to the automotive companies. This thesis investigates an approach to standardize a non-AUTOSAR AVL Hybrid Control Unit (HCU) software with AUTOSAR features.
Firstly, the deviations of the AVL HCU software from the AUTOSAR standard have been studied across the V-Model flow of Software Development Life-cycle (SDLC).
Secondly, in the conversion phase, MATLAB scripts were developed to handle the conversion and code generation process of Application Software (ASW) models to AUTOSAR format. The original Basic Software (BSW) of AVL HCU software has been reused for integration. A separate generator script has also been implemented in MATLAB for the generation of the interfacing Run Time Environment (RTE) portion.
Thirdly, in the integration phase, a total of four versions of AUTOSAR compliant HCU software with different optimization settings have been generated from the build process. In the next step, the generated AUTOSAR software versions have been evaluated in a Hardware-in-the-Loop (HIL) set-up. The Key Performance Indicators (KPIs) such as memory consumption, task runtime, CPU utilization, and stack usage, etc. have been compared for the various AUTOSAR versions with reference to the original non-AUTOSAR HCU software.
Finally, the thesis sums up the effectiveness of this direct standardization approach and discusses if a 100% AUTOSAR compatibility was achievable.