Development of the AutoSAR system model parser within the test environment generator

Automobili imaju sve više elektroničkih komponenti i složenih sustava za pomoć u vožnji. Svim ovim kompleksnim sustavima i mnogim drugim koje posjeduju današnji automobili upravljaju ugradbeni računalni sustavi visoke složenosti. Ova računala imaju velik broj komunikacijskih kanala koje je potrebno testirati, a zbog količine potrebnih testova pogodno ih je automatski generirati. U ovom diplomskom radu analizirano je postojeće rješenje generatora testnog okruženja i njegove mane, te je predloženo programsko rješenje model parsera u sklopu generatora testnog okruženja. Model parser dohvaća podatke iz .arxml datoteka i stvara bazu podataka. Bilo je potrebno ostvariti dinamičnost parsera i poboljšati performanse u odnosu na već postojeće rješenje. Model parser pisan je u programskom jeziku C++. Nakon ispunjenih zahtjeva koji su definirani za ovaj model parser provedeno je testiranje, te je utvrđeno vremensko poboljšanje od šezdeset puta u odnosu na postojeći model parser.Cars have more and more electronical components and complex systems for driver assistance in driving. These complex systems and many others, which today’s cars possess, are driven by highly complex embedded computer systems. These computers have large amounts of communication channels which need to be tested, and because of the quantity of these tests, it is suitable to generate them automatically. In this master’s thesis an existing solution of the test environment generator and its flaws are analyzed and a software solution for model parser that is a part of the test environment generator is developed. Model parser gets data from .arxml file and creates a data base. It was required to create a dynamic parser that outperforms the existing solution. Model parser was written in C++ programming language. When requirements that are defined for this model parser were met, the solution was tested. Through testing, it was shown that the new model parser is sixty times faster than the previous one