|Title:||A study on effective Test-combinations for test coverage in product-line engineering of industrial systems|
At Scania, products are based on a 'modular' design, where different type of modules (such as a type of engine) can be combined with other modules in a very large number of vehicle-configurations (individual products). Utilizing this modular system, the automotive components and features are picked up and customized to the client to an appropriate configuration (combination) and almost all vehicles that are built are different. The thesis will be carried out with the Group REST, "System and Integration Test", which is responsible for verifying implementations of distributed functionality in trucks and buses, where two or more ECUs (Electronic Control Unit) communicate with each other via the CAN (Controller Area Network). A vehicle can have 20+ ECUs connected with 3 or more different CAN-networks. During integration testing, the CAN interfaces of the control units are tested in Scania’s HIL-lab (Hardware in the loop), a 'manual' lab and in vehicles. This exhaustive testing is performed to verify that the devices work as intended. The verification work is performed in the HIL-lab, where various control functions are tested in a fully simulated environment. These automated tests offer better reproducibility and the possibility of regression testing of many configurations where each configuration is based on a real vehicle. The tests are based on user-functions such as 'cruise-control' where each test-case is realized in a test-script (if run in the lab) or a test-instruction (for vehicle tests). The results are stored in a database together with the time, configuration (vehicle-type), ECU software versions and a lot of other things that has to do with the result and environment at the specific test moment. HIL-lab cannot verify every possible configuration needed; before choosing good testing strategies and tools, first it is needed to choose the right configurations to test (both as a real vehicle and HIL configurations) and secondly evaluate the test coverage and the possible gaps in the tests.
One of the challenges is to run the tests on 'all' possible configurations and then present the results and test coverage for different vehicle types over time. The main focus for the thesis is to find methods/algorithms on how to choose which complete vehicle-configurations to run the tests on. Depending on time and interest other parts of the work could be to find strategies for how to run our tests efficiently until a complete coverage of all the needed modules/configurations are obtained. The work includes study of the Scania specification and modular-system and the selection strategies used today. It will also include studying of different techniques for test and test-selection as well as obtaining an overall understanding of vehicle electronics such as CAN-networks and user distributed functions. Software development of selection-tools for demo of the algorithms using C#/.NET/SQL-server and/or Python is also envisaged.
|Company:||SCANIA CV AB, kontaktperson: Pär Sundbäck|
|Prel. end date:||30-10-2010|
|Student:||Rabia Durrani email@example.com|