Primjena strojnog učenja za upravljanje raskrižjima u okruženju mješovitih prometnih tokova

master's thesis

• Title (english): Application of Machine Learning for Intersection Control in Mixed Flow Environment
• Author: Žana Jakobović
• Mentor: Edouard Ivanjko (mentor)
• Mentor: Dino Čakija (komentor)
• Committee member: Miroslav Vujić (predsjednik povjerenstva)
• Committee member: Martin Gregurić (član povjerenstva)
• Committee member: Sadko Mandžuka (član povjerenstva)
• Granter: University of Zagreb; Faculty of Transport and Traffic Sciences; (Division of Intelligent Transport Systems and Logistics); (Department of Intelligent Transport Systems); Zagreb
• Defense date and country: 2021-07-06, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Traffic and Transport Technology; Intelligent Transport Systems and Logistics

Abstract

Porast broja stanovništva uzrokovao je povećanje prijevozne potražnje što dovodi do svakodnevnih prometnih zagušenja, posebno u vršnim satima. Problem nastaje kada prijevozna potražnja postane veća od prijevozne ponude te prometnice nemaju dovoljan kapacitet za visoku razinu usluge. U gradskoj sredini zagušenja su izražena oko semaforiziranih raskrižja jer dolazi do preplitanja dvaju ili više prometnih tokova. Kako bi se smanjilo vrijeme putovanja i duljine repova čekanja moguće je koristiti metode strojnog učenja za prilagodbu signalnih planova. Pri tome mješoviti prometni tokovi koji sadrže ljudski upravljana vozila i umrežena autonomna vozila predstavljaju novi nadolazeći upravljački problem. Problemi nastaju zbog različitog percipiranja okoline i poštivanja pravila ponašanja kod vozača u ljudski upravljanim vozilima te autonomnim i umreženim autonomnim vozilima. U radu su opisane značajke mješovitih prometnih tokova, upravljački koncepti za slučaj semaforiziranih raskrižja i mješovitih tokova te analizirane mogućnosti primjene metoda strojnog učenja za upravljanje mješovitim prometnim tokovima.

Abstract (english)

The increase of the population caused the increase in transport demand which causes daily traffic congestions, especially during peak hours. The problem arises when traffic demand exceeds the traffic offer, and roads do not have enough capacity for a high level of service. In the urban environment, congestions are especially expressed around signalized intersections because two or more traffic flow intertwine. To decrease the travel time and waiting queue length, it is possible to implement machine learning methods to adapt intersection signal plans. Thereby, mixed traffic flows consisting of human-driven vehicles and connected autonomous vehicles represent an upcoming control issue. Problems arise due to different environment perceptions and respecting the rules with human-driven vehicles and autonomous and connected autonomous vehicles. In this thesis, mixed traffic flow features, traffic management concepts for signalized intersections, mixed traffic flows are described, and the possibilities of implementing machine learning methods for managing mixed traffic flows are analyzed.

• Keywords: mješoviti prometni tok; semaforizirana raskrižja; strojno učenje; umrežena autonomna vozila; autonomna vozila; umjetna inteligencija
• Keywords (english): mixed traffic flow; signalized intersection; machine learning; connected autonomous vehicles; autonomous vehicles; artificial intelligence
• Language: croatian
• URN:NBN: urn:nbn:hr:119:997364
• Study programme: Title: Intelligent Transport Systems and Logistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka prometa (magistar/magistra inženjer/inženjerka prometa)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2021-07-09 14:21:36