| Titre : | Autonomous Robot Navigation in Dense Forests Using Reinforcement Learning |
| Auteurs : | NEDJMA HACHEMI, Auteur ; Ahmed Tibermacine, Directeur de thèse |
| Editeur : | Biskra [Algérie] : Faculté des Sciences Exactes et des Sciences de la Nature et de la Vie, Université Mohamed Khider, 2024 |
| Format : | 1 vol. (94 p.) / ill., couv. ill. en coul / 30 cm |
| Langues: | Français |
| Résumé : |
Cette th`esepr´esenteuneapprochenovatricepourlanavigationautonomedesrobotsdansdesenviron- nementscomplexes,avecunaccentparticuliersurlesforˆetsdenses,etdesapplicationsdanslasurveil- lance environnementale,lagestionforesti`ere,leso´erationsderechercheetsauvetage,etl’agriculture. Le mod`elepropos´eint`egreMaskR-CNNpouruneperceptionvisuelleavanc´ee,permettantaurobot de d´etecteretsegmenterdescheminsnavigablesdansdesterrainsdifficiles,ainsiquel’algorithme d’apprentissageparrenforcementSoftActor-Critic(SAC)pourunepriseded´ecisionadaptative.Ce syst`emehybridepermetaurobotd’apprendredesstrat´egiesdemouvementoptimalestoutennaviguant dans desenvironnementsdynamiques,en´evitantlesobstaclesetengarantissantlas´ecurit´e.Entraˆın´e dans desconditionssimul´eesdeforˆet,lemod`eled´emontreuneperformancerobusteensegmentation de cheminentempsr´eeletennavigationadaptative.Larecherchecontribue`alarobotiqueautonome en am´eliorantlacapacit´edesrobots`apercevoirets’adapter`adesenvironnementsnonstructur´es, avecdesperspectivesfuturesexplorantl’int´egrationdecapteursmultimodaux,larobotiqueenessaim, l’apprentissageentempsr´eeletlaplanificationdetrajets´economesen´energie. ii |
| Sommaire : |
Contents List ofFigures viii List ofTables x General Introduction 1 Chapter 1:Robotics 3 1 Introduction 3 2 Robots 3 2.1 Appearance............................................3 2.2 Brain...............................................3 2.3 Sensors..............................................4 2.4 Actuators.............................................5 2.5 Program..............................................5 2.6 Behavior..............................................5 3 Typesofrobots 5 3.1 WheeledRobots.........................................5 3.2 LeggedRobots..........................................5 3.3 AerialRobots...........................................6 3.4 AquaticRobots..........................................7 3.5 CrawlerRobots..........................................7 3.6 HumanoidRobots........................................8 3.7 Snake-likeRobots.........................................8 3.8 Exoskeletons...........................................9 3.9 Nanorobots............................................9 4 Roboticsapplications10 4.1 Security..............................................10 4.2 Entertainment..........................................11 4.3 SpaceExploration........................................11 4.4 Agriculture............................................11 4.5 HealthCare............................................12 4.6 UnderwaterExploration.....................................12 4.7 Manufacturing..........................................13 4.8 CustomerService.........................................13 4.9 Military..............................................13 4.10 Education............................................14 5 RoboticTasks 14 5.1 Manipulation...........................................14 5.2 MovementandNavigation...................................15 5.3 SensingandPerception.....................................15 6 ChallengesforRobotsinDiverseEnvironments16 6.1 UrbanEnvironment.......................................16 6.2 UnderwaterEnvironment....................................16 6.3 ForestEnvironment.......................................17 7 Conclusion 17 Chapter 2:Artificialintelligenceinrobotics18 1 Introduction 18 2 Artificialintelligence18 3 ArtificialIntelligence:EmpoweringRoboticControl18 3.1 NarrowAI.............................................18 3.2 MachineLearning(ML):LearningfromExperience......................19 3.3 Computervision.........................................21 3.4 NaturalLanguageProcessing(NLP)..............................22 3.5 Sensorfusion...........................................23 3.6 DeepLearning..........................................23 4 ApplicationofAIinrobotics25 4.1 Perception.............................................25 4.2 Decisionmaking.........................................25 4.3 Control..............................................25 4.4 Interactionwiththeenvironment...............................26 5 Reinforcementlearninganditsroleinroboticnavigation26 5.1 LearningNavigationPolicieswithRL.............................26 5.2 ComponentsofRL-basedNavigationSystems.........................26 5.3 RLAlgorithmsforRoboticNavigation.............................27 5.4 ChallengesandConsiderations.................................28 5.4.1 ApplicationsofRL-basedNavigation..........................28 6 ChallengesinRoboticNavigation29 6.1 High-DimensionalStateSpace..................................29 6.2 PartialObservability.......................................29 6.3 DynamicEnvironments.....................................30 6.4 SafetyConstraints........................................30 6.5 Long-TermPlanning.......................................30 6.6 AdaptabilitytoNovelEnvironments..............................30 6.7 Human-RobotInteraction....................................30 7 Conclusion 30 Chapter 3:ProposedModel:DesignandDevelopment32 1 Introduction 32 2 GeneralDesignoftheProposedModel32 3 MaskR-CNN-BasedVisualPerception34 3.1 MaskR-CNNArchitectureandFlowchartExplanation....................34 3.2 Step-by-StepBreakdownofMaskR-CNNComponents:...................35 3.3 WhyMaskR-CNNforTrailDetectioninForests:.......................36 4 Dataset 37 5 RobotDescriptionandNavigationMethodology38 5.1 RobotDescription........................................38 5.2 NavigationMethodology.....................................39 5.2.1 PathPredictionusingRCNN..............................39 5.2.2 ObstacleAvoidanceusingProximitySensors.....................39 5.2.3 ControlSystem:PurePursuitController.......................40 5.3 IntegratedNavigationSystem..................................40 5.4 TestingEnvironment.......................................41 6 SAC-BasedRobotNavigationandExploration43 6.1 ProblemFormulation.......................................43 6.2 BackgroundonSoftActor-CriticAlgorithm..........................44 6.2.1 Entropy-AugmentedObjective.............................44 6.2.2 PolicyandValueFunctions...............................44 6.2.3 PolicyUpdate......................................44 6.2.4 Q-FunctionUpdate....................................45 6.2.5 AutomaticEntropyAdjustment.............................45 6.3 SACNeuralNetworkArchitectures...............................45 6.3.1 Overview.........................................45 6.3.2 FusionModel.......................................45 6.3.3 PolicyNetwork......................................46 6.3.4 CriticNetworks......................................46 6.3.5 InitializationandActivationFunctions.........................46 6.3.6 NetworkArchitectures..................................47 6.4 TrainingAlgorithmandRewardFunction...........................48 6.4.1 ExperienceCollectionandReplayBuffer........................48 6.4.2 NetworkUpdates.....................................48 6.4.3 RewardFunctionDesign.................................48 7 Software 49 8 Results 55 General Conclusion 59 References 61 |
| Type de document : | Mémoire master |
Disponibilité (1)
| Cote | Support | Localisation | Statut |
|---|---|---|---|
| MINF/918 | Mémoire master | bibliothèque sciences exactes | Consultable |
