Korekcija navika vozača strojnim učenjem

undergraduate thesis

• Title (english): Correction of driver habits with machine learning
• Author: Jan Kuzman
• Mentor: Vinko Lešić (mentor)
• Committee member: Vinko Lešić (član povjerenstva)
• Committee member: Filip Vrbanc (član povjerenstva)
• Committee member: Mihael Jakšić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-07-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Potrošnja goriva u vozilima je pod velikim utjecajem stila vožnje, koji može varirati od sporog do agresivnog. Koristeći strojno učenje nad prikupljenim podacima iz akcelerometra i žiroskopa, ovaj rad istražuje kako se potrošnja goriva mijenja ovisno o stilu vožnje. Razvijen je model za predikciju potrošnje goriva korištenjem algoritma Random Forest Regression. Podaci su prikupljeni iz senzora ugrađenih u mobilnim uređajima, a obuhvaćaju razdoblje vožnje te su podijeljeni u tri kategorije – agresivna, normalna i spora vožnja. Model je treniran na segmentiranim podacima, a njegova točnost je testiranja na odvojenom i još neviđenom skupu podataka. Postignuta je visoka, skoro savršena, točnost za sve stilove vožnje. Uspješno su implementirani sustavi upozorenja i nagrađivanja vozača koji mogu pomoći u poticanju ekonomičnije vožnje, smanjujući potrošnju goriva i emisije stakleničkih plinova.

Abstract (english)

Fuel consumption in vehicles is greatly influenced by driving style, which can vary from slow to aggressive. Utilizing machine learning on collected data from accelerometers and gyroscopes, this paper investigates how fuel consumption changes depending on the driving style. A model for predicting fuel consumption was developed using the Random Forest Regression algorithm. The data was collected from sensors embedded in mobile devices, covering the driving period and categorized into three categories – aggressive, normal and slow driving. The model was trained on segmented data, and its accuracy was tested on separate and previously unseen dataset. High, nearly perfect accuracy was achieved for all driving styles. Successfully implemented driver warning and reward system can help promote more economical driving, reducing fuel consumption and greenhouse gas emissions.

• Keywords: strojno učenje; Random Forest Regression; predikcija potrošnje goriva; korekcija navika
• Keywords (english): machine learning; Random Forest Regression; fuel consumption prediction; habit correction
• Language: croatian
• URN:NBN: urn:nbn:hr:168:945348
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-10-12 21:16:26