Simulacija autonomnih vozila u CARLA programskom okviru

undergraduate thesis

• Title (english): Simulation of Autonomous Vehicles in CARLA Simulator
• Author: Leonardo Žalar
• Mentor: Vladimir Čeperić (mentor)
• Committee member: Vladimir Čeperić (predsjednik povjerenstva)
• Committee member: Josip Šarić (član povjerenstva)
• Committee member: Ivan Grubišić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-09-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Područje autonomnih vozila bilo je veoma zanimljivo za mnoge istraživačke skupine i proizvođače automobila te ima dugu povijest razvoja. Izgradanja prototipova autonomnih vozila je skupa te nije moguće u stvarnom svijetu prototip isprobati u mnogim situacijama zato što bi bile preopasne. Stoga, simulatori za autonomna vozila dopuštaju nam brži i jeftiniji razvoj te omogućuju podvrgavanje autonomnih agenata situacijama koje su preopasne u stvarnosti. CARLA je simulator urbane vožnje otvorenog koda izgrađen za razvoj, treniranje te validaciju autonomnih vozila. CARLA simulator ima velik broj ugrađenih senzora. Kako bi simulirali autonomnu vožnju trenirana su 3 različita agenta dubokim Q-učenjem u CARLA simulatoru te svaki koristi drugačiji senzor za navigiranje okolinom. Od senzora korišteni su RBG kamera, radar i lidar.

Abstract (english)

The field of autonomous vehicles has been very interesting for many research groups and car manufactures for a long time and has a rich history of development. Building autonomous vehicle prototypes is costly and it's not possible for the prototype to be tested under all circumstances because it would be too dangerous. Autonomous vehicle simulators allow for faster and cheaper development of autonomous vehicles and enable us to test our simulated car in situations that would be too dangerous in real life. CARLA is an open urban driving simulator built for the development, training and validation of autonomous vehicles. CARLA has a lot of in-built sensors. In order to simulate autonomous driving, 3 different agents have been trained with deep Q-learning in the CARLA simulator and each agent used a different sensor to navigate in the environment, the sensors that have been used are: RGB camera, lidar and radar.

• Keywords: CARLA; simulacija; duboko učenje; Q-učenje; potporno učenje; strojno učenje; autonomna vožnja; autonomni auto; lidar; radar; RGB kamera; GTA V; samovozeći automobil
• Keywords (english): CARLA; simulation; deep learning; Q-learning; reinforcement learning; machine learning; autonomous driving; autonomous car; lidar; radar; RGB camera; GTA V; self-driving car
• Language: croatian
• URN:NBN: urn:nbn:hr:168:973119
• 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: 2023-01-14 10:20:53