Αυτοματοποιημένη οδήγηση οχημάτων σε περιβάλλον προσομοίωσης με χρήση τεχνικών ενισχυόμενης μάθησης

Χρυσικόπουλος, Γεώργιος

Automated vehicle driving using reinforcement learning techniques in simulated environments

Bachelor Dissertation

Author

Χρυσικόπουλος, Γεώργιος

Date

2025-09

Abstract

This thesis aims to develop an autonomous driving agent in a virtual simulation environment using reinforcement learning techniques. Initially, a literature search was carried out with the aim of gaining a deeper understanding of the relevant technologies, the algorithms used in reinforcement learning, as well as the CARLA simulation environment (version 0.9.15), which was chosen for the implementation and evaluation of the agent. The algorithm implemented is the Double Deep Q-Network (DDQN), an improved reinforcement learning algorithm that allows the agent to learn optimal driving policies through interaction with the environment. The work focuses on the challenges that arise even in controlled, simulated environments, highlighting both the possibilities and limitations of these algorithms in autonomous vehicle training. By presenting the experiments and results, this research contributes to a broader understanding of the application and scalability of reinforcement learning methods in the field of automated driving systems.

Department

Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Πληροφορικής

Number of pages

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/18147

Collections

Τμήμα Πληροφορικής

Show full item record

Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα