Institutional Repository [SANDBOX]
Technical University of Crete
EN  |  EL

Search

Browse

My Space

Reinforcement learning algorithms in autonomous driving, a comparitive evaluation

Matsioris Georgios

Full record


URI: http://purl.tuc.gr/dl/dias/0213210D-6E8B-42AB-91D3-EA9C93848998
Year 2024
Type of Item Diploma Work
License
Details
Bibliographic Citation Georgios Matsioris, "Reinforcement learning algorithms in autonomous driving, a comparitive evaluation", Diploma Work, School of Production Engineering and Management, Technical University of Crete, Chania, Greece, 2024 https://doi.org/10.26233/heallink.tuc.100493
Appears in Collections

Summary

The objective of this thesis is to develop a methodological framework that facilitates the development of autonomous agents for guiding electric vehicles with autonomous navigation capabilities. To develop this framework, we utilize the model of an urban electric car equipped with a variety of sensors. Initially, a simulated model from a previous study is enhanced. The functionality of this model is tested in a simulated environment based on the CARLA software using a specially designed GYM environment. Subsequently, autonomous agents are developed to guide the vehicle model along a predetermined path in an urban setting. This process relies on the A* algorithm to generate the desired trajectory, followed by using the corresponding data to train the autonomous agents. Two different reinforcement learning algorithms are employed to train the agents, and a comparative study of the results is conducted.

Available Files

Services

Statistics