Centre de documentation
scientifique

Exploration of reinforcement learning algorithms for autonomous vehicle visual perception and control / Florence Carton (2021)  

Public Titre : Exploration of reinforcement learning algorithms for autonomous vehicle visual perception and control Titre original : Exploration des algorithmes d'apprentissage par renforcement pour la perception et le controle d'un véhicule autonome par vision Type de document : Thèse/HDR Auteurs : Florence Carton, Auteur ; David Filliat, Directeur de thèse Editeur : Paris : Ecole Nationale Supérieure des Techniques Avancées ENSTA Année de publication : 2021 Importance : 173 p. Format : 21 x 30 cm Note générale : bibliographie Thèse de Doctorat de l’Institut Polytechnique de Paris, Spécialité : Informatique, Données, IA Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Intelligence artificielle [Termes IGN] apprentissage par renforcement [Termes IGN] classification dirigée [Termes IGN] instrument embarqué [Termes IGN] navigation autonome [Termes IGN] reconnaissance de formes [Termes IGN] réseau neuronal profond [Termes IGN] robot mobile [Termes IGN] segmentation sémantique [Termes IGN] vision par ordinateur Index. décimale : THESE Thèses et HDR Résumé : (auteur) Reinforcement learning is an approach to solve a sequential decision making problem. In this formalism, an autonomous agent interacts with an environment and receives rewards based on the decisions it makes. The goal of the agent is to maximize the total amount of rewards it receives. In the reinforcement learning paradigm, the agent learns by trial and error the policy (sequence of actions) that yields the best rewards.In this thesis, we focus on its application to the perception and control of an autonomous vehicle. To stay close to human driving, only the onboard camera is used as input sensor. We focus in particular on end-to-end training, i.e. a direct mapping between information from the environment and the action chosen by the agent. However, training end-to-end reinforcement learning for autonomous driving poses some challenges: the large dimensions of the state and action spaces as well as the instability and weakness of the reinforcement learning signal to train deep neural networks.The approaches we implemented are based on the use of semantic information (image segmentation). In particular, this work explores the joint training of semantic information and navigation.We show that these methods are promising and allow to overcome some limitations. On the one hand, combining segmentation supervised learning with navigation reinforcement learning improves the performance of the agent and its ability to generalize to an unknown environment. On the other hand, it enables to train an agent that will be more robust to unexpected events and able to make decisions limiting the risks.Experiments are conducted in simulation, and numerous comparisons with state of the art methods are made. Note de contenu : 1- Introduction 2- Supervised learning and reinforcement learning background 3- State of the art 4- End-to-end autonomous driving on circuit with reinforcement learning 5- From lane following to robust conditional driving 6- Exploration of methods to reduce overfit 7- Conclusion Numéro de notice : 28325 Affiliation des auteurs : non IGN Thématique : INFORMATIQUE Nature : Thèse étrangère Note de thèse : Thèse de Doctorat : Informatique, Données, IA : ENSTA : 2021 DOI : sans En ligne : https://tel.hal.science/tel-03273748/ Format de la ressource électronique : URL Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=98363