Simple Task Scheduler in Embedded Soft Real-Time Systems

U okviru ovog rada razvijen je algoritam za raspoređivanje zadataka u sustavima za rad stvarnom vremenu s blagim vremenskim ograničenjima. Cilj je poboljšati iskoristivost procesorskog vremena i time bolje iskoristiti sklopovlje, kao i povećati robusnost sustava, pritom zadržavajući svojstva sustava za rad u stvarnom vremenu. Algoritam je testiran korištenjem statički i dinamički generiranih primjera u simulaciji, kao i na stvarnom sustavu Elektroničke kontrolne jedinice automobila. Za ispitivanje rezultata korišteni su metrički pokazatelji. Rezultati su se pokazali zadovoljavajući. Sustavi koji su razmatrani prilikom stvaranja algoritma su ugradbeni sustavi koji iterativno poboljšavaju rezultate, poput primjerice algoritama umjetne inteligencije, genetskih algoritama, heurističkih algoritama i algoritama interpretacije signala.Within this paper, a task scheduler algorithm was developed for real-time working systems with mild time constraints. The goal is to improve the usability of processor time and thus make better use of the hardware as well as increase system robustness while retaining the system's real-time performance features. The algorithm was tested with statically and dynamically generated examples, as well as on the Electronic Control Unit of an automotive vehicle. Metric indicators were used for the evaluation of the results. The results were satisfactory. The systems considered in creating the algorithm are embedded systems that iteratively improve the results, such as artificial intelligence algorithms, genetic algorithms, heuristic algorithms and signal interpretation algorithms