20.500.12556/DKUM-98305 Uporaba umetne inteligence za ugotavljanje vpliva voznika na porabo energije za pogon osebnega vozila magistrsko delo Using artificial intelligence to determine the driver's impact on energy consumption in personal vehicles V magistrskem delu je obravnavano napovedovanje porabe goriva na podlagi podatkov, zbranih med realno vožnjo z vmesnikom OBD-II in senzorji mobilnega telefona. Meritve so bile zbrane z enim voznikom in enim osebnim vozilom z bencinskim motorjem s prisilnim polnjenjem na 29 vožnjah. Podatki so bili predobdelani in uporabljeni za učenje regresijskih modelov strojnega učenja. Najboljše rezultate je dosegel naključni gozd z izborom značilk ElasticNet. Rezultati so pokazali, da so za napoved pomembne predvsem značilke hitrosti, položaja dušilne lopute, obremenitve motorja in pospeševanja. Na osnovi modela je bil izveden osnovni lokalni priporočilni sistem za varčnejšo vožnjo. This master’s thesis investigates fuel consumption prediction using data collected in real-world driving from an on-board diagnostics interface and smartphone sensors. Measurements were collected over 29 drives by one driver in one passenger vehicle with a forced-induction gasoline engine. The data was cleaned, segmented into time windows, and transformed into driving-related features. Several regression-based machine learning models were evaluated, with the best performance achieved by a Random Forest model combined with ElasticNet feature selection. The results indicate that speed, throttle position, engine load, and acceleration-related features are important predictors. The selected model also supported an offline eco-driving recommendation system. poraba goriva OBD-II strojno učenje vozni slog priporočilni sistem fuel consumption OBD-II machine learning driving style recommendation system true true false [L. Zorman] Slovenski jezik Angleški jezik Magistrsko delo/naloga 2026-06-03 21:37:52 2026-06-29 12:41:52 2026-07-04 03:09:45 0000-00-00 00:00:00 2026 0 Maribor 0 Maribor 1 spletni vir (1 datoteka PDF (IX, 94 str.)) 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.85.021:519.2(043.2) 283569923 MAG_Zorman_Lina_2026.pdf MAG_Zorman_Lina_2026.pdf 1 61A1E863489EACCD757EBB482E63C0E6 e50c587a1396a35242ff5277f1e2c8745456aeaf874019619f95814db0991744 cddfbc09-5f84-11f1-8a3a-00155d580e34 https://dk.um.si/Dokument.php?lang=slv&id=206723 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0