학위논문 (석사)-- 서울대학교 대학원 : 전기공학부, 2013. 8. 오성회. ; Gaussian process is a powerful probabilistic estimation tool which is used widely in engineering fields such as Computer vision, Robotics and sensor networks, etc. This thesis implemented an estimation algorithm of the total map with sparse sensing data using Gaussian Process. In the implemented algorithm, two kinds of kernel functions are applied to the spatial Gaussian Process model ; squared exponential kernel and neural network kernel. The performance of the proposed algorithm was verified by the experiments with a simple mobile sensor network. To construct a simple mobile sensor network based on ROS (Robot Operating System) platform, a two wheeled mobile robot (Pioneer3DX) and a two dimensional laser scanner (SICKlms200) are used. ; 1 Introduction 1 1.1 Mobile Sensor Network 1.2 Simulataneous Localization and Mapping (SLAM) 1.3 Occupancy Grid Map 2 Related Work 2.1 Mapping 2.2 Sensing and Locating 2.3 Probabilistic Solution for Mapping Problem 3 Gaussian Process (GP) 3.1 Weight-space View 3.1.1 The Standard Linear Model 3.1.2 Projections of Inputs into Feature Space 3.2 Function-space View 3.2.1 Prediction with Noise-free Observations 3.2.2 Prediction using Noisy Observations 3.3 Varying Hyperparameters 4 GP Applied to Mapping Problem 4.1 Overview of Contextual Mapping 4.2 Training Hyperparameters 5 Experimental results 5.1 The Mapping Problem in a Real Indoor Environment 5.2 GP Estimation for Single Frame of Laser Scanner 5.2.1 The given Training Data 5.2.2 Selecting a Kernel Function 5.2.3 Optimizing Hyper-parameters 5.2.4 Estimation 5.3 The Estimation Problem in a Simulated Environment 6 Conclusion 6.1 Contribution of GP for Mapping Problem 6.2 Future Works ; Master


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Probabilistic Estimation of Incomplete Map Using Gaussian Process



    Erscheinungsdatum :

    2013-01-01


    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629



    Probabilistic Lane Change Prediction using Gaussian Process Neural Networks

    Kruger, Martin / Novo, Anne Stockem / Nattermann, Till et al. | IEEE | 2019


    Augmented tracking with incomplete observation and probabilistic reasoning

    Xu, M. / Ellis, T. | British Library Online Contents | 2006


    Probabilistic Integration of Material Process Modeling and Fracture Risk Assessment Using Gaussian Process Models

    Enright, M.P. / McFarland, J. / McClung, R. et al. | British Library Conference Proceedings | 2013


    Probabilistic Integration of Material Process Modeling and Fracture Risk Assessment Using Gaussian Process Models

    Enright, Michael P. / McFarland, John / McClung, Robert et al. | AIAA | 2013