@inproceedings{SeidlKraemerMottok, author = {Seidl, M. and Kr{\"a}mer, Stefan and Mottok, J{\"u}rgen}, title = {Improving the reliability of an embedded real-time system by applying coded processing}, series = {4th Applied Research Conference, 5th July 2014, Ingolstadt}, booktitle = {4th Applied Research Conference, 5th July 2014, Ingolstadt}, publisher = {Shaker}, address = {Aachen}, isbn = {978-3-8440-2875-1}, pages = {70 -- 74}, language = {en} } @inproceedings{MuchaMottokKraemer, author = {Mucha, Matthias and Mottok, J{\"u}rgen and Kr{\"a}mer, Stefan}, title = {Estimation of Worst Case Response Time Boundaries in Multi-Core Real-Time Systems}, series = {2017 International Conference on Applied Electronics (AE), 5-6 Sept. 2017, Pilsen}, booktitle = {2017 International Conference on Applied Electronics (AE), 5-6 Sept. 2017, Pilsen}, publisher = {IEEE}, isbn = {978-80-261-0641-8}, issn = {1803-7232}, doi = {10.23919/ae.2017.8053598}, pages = {1 -- 6}, abstract = {We address a novel probabilistic approach to estimate the Worst Case Response Time boundaries of tasks. Multi-core real-time systems process tasks in parallel on two or more cores. Tasks in our contribution may preempt other tasks, block tasks with semaphores to access global shared resources, or migrate to another core. The depicted task behavior is random. The shape of collected response times of a task within a processing time is multimodal. Extreme Value approaches need unimodal response time distributions to estimate the Worst Case Response Time of tasks. The new proposed method derives a set of three task set shapes from the source task set. It is used to minimize the uncertainty of random task behavior by maximizing the coverage of possible Worst Case Response Times. The case study evaluates the new proposed estimation method by the use of dynamically generated random tasks with varying task properties.}, language = {en} }