de
en
Schliessen
Detailsuche
Bibliotheken
Projekt
Impressum
Datenschutz
zum Inhalt
Detailsuche
Schnellsuche:
OK
Ergebnisliste
Titel
Titel
Inhalt
Inhalt
Seite
Seite
Im Dokument suchen
Rolf, Matthias: Goal babbling for an efficient bootstrapping of inverse models in high dimensions. 2012
Inhalt
Contents
List of Figures
Introduction
Motivation
Outline
Autonomous Learning of Coordination Skills
The Coordination Problem
Internal Models for Coordination
Exhaustive Learning of Forward Models
Learning of Inverse Models
A Framework for Goal Babbling
Inspiration from Infant Development
Concept: Goal Babbling
Method: Learning Inverse Models from Examples
Inversion of Causality in Linear Domains
Two Spaces and their Gradients in Linear Domains
Fixpoint Analysis for Explorative Learning
Plain Goal-Directed Exploration
Exploratory Noise
Numeric simulation results
Discussion
Coordination Problems with Non-Convex Solution Sets
Non-convex Solution Sets during Goal Babbling
Structured Variation and Regularization
Examples
Experiments
Planar Arm: 1D Coordination Task
Planar Arm: 2D Coordination Task
Humanoid robot: 3D Coordination Task
Discussion
Online Learning Dynamics during Goal Babbling
Online Learning in the Loop
Online Goal Babbling Formulation
Continuous Path Generation
Structured Continuous Variation
Incremental Regression Model
Experiments
Effects of the Learning Rate
Scalability
Discussion
Application on a Bionic Elephant Trunk
Bionic Handling Assistant Setup
Actuation and Sensing
Accuracy and Limits
Kinematic Coordination Problem
Online Goal Babbling Formulation
BHA Experiments
Learning to Reach on the BHA
Local Error Correction
Non-Stationary Behavior in Simulation
Kinematic Simulation of the Bionic Handling Assistant
Varying Ranges and Sensory Drifts
Morphological Growth
Discussion
Conclusion
Summary
Discussion
Outlook
Bibliography