Publicação
Artificial intelligence-based control of autonomous vehicles in simulation: a CNN vs. RL case study
| datacite.subject.fos | Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática | |
| datacite.subject.fos | Engenharia e Tecnologia::Engenharia Mecânica | |
| datacite.subject.sdg | 04:Educação de Qualidade | |
| datacite.subject.sdg | 11:Cidades e Comunidades Sustentáveis | |
| datacite.subject.sdg | 09:Indústria, Inovação e Infraestruturas | |
| dc.contributor.author | Vasiljević, Ive | |
| dc.contributor.author | Musić, Josip | |
| dc.contributor.author | Lima, José | |
| dc.date.accessioned | 2026-03-02T15:13:21Z | |
| dc.date.available | 2026-03-02T15:13:21Z | |
| dc.date.issued | 2024 | |
| dc.description.abstract | The article provides a comparison of Convolutional Neural Network (CNN) and Reinforcement Learning (RL) applied to the field of autonomous driving within the CARLA (CAr Learning to Act) simulator for training and evaluation. The analysis of results revealed CNNs better overall performance, as it demonstrated a more refined driving experience, shorter training durations, and a more straightforward learning curve and optimization process. However, it required data labelling. In contrast, RL relayed on an exhaustive (unsupervised) exploration of different models, ultimately selecting the model at timestep 600,000, which had the highest mean reward. Nevertheless, RL’s approach revealed its susceptibility to excessive oscillations and inconsistencies, necessitating additional optimization and tuning of hyperparameters and reward functions. This conclusion is further substantiated by a range of used performance metrics (objective and subjective), designed to assess the performance of each approach. | eng |
| dc.identifier.citation | Vasiljević, Ive; Musić, Josip; Lima, José (2024). Artificial intelligence-based control of autonomous vehicles in simulation: a CNN vs. RL case study. In Second International Conference, MoStart 2024. Cham: Springer Nature, p. 124-151. ISBN 978-3-031-62058-4. DOI: 10.1007/978-3-031-62058-4_10 | |
| dc.identifier.doi | 10.1007/978-3-031-62058-4_10 | |
| dc.identifier.isbn | 978-3-031-62057-7 | |
| dc.identifier.isbn | 978-3-031-62058-4 | |
| dc.identifier.uri | http://hdl.handle.net/10198/35916 | |
| dc.language.iso | eng | |
| dc.peerreviewed | yes | |
| dc.publisher | Springer Nature | |
| dc.relation.ispartof | Communications in Computer and Information Science | |
| dc.relation.ispartof | Digital Transformation in Education and Artificial Intelligence Application | |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | |
| dc.subject | Reinforcement learning | |
| dc.subject | CNN | |
| dc.subject | CARLA simulator | |
| dc.title | Artificial intelligence-based control of autonomous vehicles in simulation: a CNN vs. RL case study | por |
| dc.type | conference paper | |
| dspace.entity.type | Publication | |
| oaire.citation.endPage | 151 | |
| oaire.citation.startPage | 124 | |
| oaire.citation.title | Second International Conference, MoStart 2024 | |
| oaire.version | http://purl.org/coar/version/c_ab4af688f83e57aa | |
| person.familyName | Lima | |
| person.givenName | José | |
| person.identifier | R-000-8GD | |
| person.identifier.ciencia-id | 6016-C902-86A9 | |
| person.identifier.orcid | 0000-0001-7902-1207 | |
| person.identifier.rid | L-3370-2014 | |
| person.identifier.scopus-author-id | 55851941311 | |
| relation.isAuthorOfPublication | d88c2b2a-efc2-48ef-b1fd-1145475e0055 | |
| relation.isAuthorOfPublication.latestForDiscovery | d88c2b2a-efc2-48ef-b1fd-1145475e0055 |