Adaptivno upravljanje izoliranim semaforiziranim raskrižjem primjenom neuronskih mreža

master's thesis

• Title (english): Adaptive Control of Isolated Signalized Intersection using Neural Networks
• Author: Mladen Miletić
• Mentor: Edouard Ivanjko (mentor)
• Committee member: Sadko Mandžuka (predsjednik povjerenstva)
• Committee member: Miroslav Vujić (član povjerenstva)
• Committee member: Luka Novačko (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: 2019-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Traffic and Transport Technology; Intelligent Transport Systems and Logistics

Abstract

Zbog sve većeg rasta prometne potražnje u gradskim sredinama dolazi do značajnog smanjenja razine uslužnosti mobilnosti u gradskim sredinama. Dio problema su i zastarjeli sustavi upravljanja semaforiziranim raskrižjem zasnovani na ustaljenim signalnim planovima. Razvojem sustava zasnovanih na umjetnoj inteligenciji moguće je riješiti neke od problema upravljanja semaforiziranim raskrižjem. Korištenjem umjetne neuronske mreže i algoritma ojačanog učenja moguće je izgraditi sustav za adaptivno upravljanje semaforiziranim raskrižjem. U ovom radu je simulirano i analizirano izolirano semaforizirano raskrižje koristeći razvijeno simulacijsko okruženje uz algoritam ojačanog učenja zasnovanog na neuronskoj mreži. Iz rezultata se zaklju čuje da se korištenjem neuronskih mreža u kombinaciji s ojačanim učenjem mogu postići poboljšanja u upravljanju semaforiziranog raskrižja.

Abstract (english)

Due to increasing traffic demand in urban areas there is a significant decrease od the level of service of urban mobility. Part of the problem are old control systems of signalized intersection that use fixed signal programs. By developing systems based on artificial intelligence it is possible to solve some of the problems of signalized intersection control. With the use of artificial neural networks and reinforcement learning algorithms it is possible to create a system for adaptive control of a signalized intersection. In this thesis, an isolated signalized intersection is simulated and analyzed with the implementation of reinforcement learning algorithm based on neural networks. From the results it can be concluded that the use of neural networks and reinforcement learning algorithms can achieve improvements in signalized intersection control.

• Keywords: Inteligentni transportni sustavi; semaforizirano raskrižje; signalni plan; neuronske mreže; ojačano učenje
• Keywords (english): Intelligent Transport Systems; Signalized Intersection; Signal Program; Neural Networks; Reinforcement Learning
• Language: croatian
• URN:NBN: urn:nbn:hr:119:502434
• 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: 2019-07-16 10:06:00