de
en
Schliessen
Detailsuche
Bibliotheken
Projekt
Impressum
Datenschutz
zum Inhalt
Detailsuche
Schnellsuche:
OK
Ergebnisliste
Titel
Titel
Inhalt
Inhalt
Seite
Seite
Im Dokument suchen
Hemion, Nikolas: Building Blocks for Cognitive Robots: Embodied Simulation and Schemata in a Cognitive Architecture. 2013
Inhalt
1 Introduction
1.1 Research Goals and Contributions of this Thesis
1.2 Outline
2 Cognitive Architecture: Overview of Theoretical Paradigms and Computational Models
2.1 Structure of the Cortex: A Brief Introduction
2.2 Cognitivism
2.2.1 Computational Models
2.2.2 Hybrid Architectures
2.2.3 Implications
2.3 Behavior-based Robotics
2.3.1 Computational Models
2.3.2 Implications
2.4 Connectionism
2.4.1 Computational Models
2.4.2 Implications
2.5 Dynamicism
2.5.1 Dynamic Field Theory
2.5.2 Computational Models
2.5.3 Implications
2.6 Discussion
3 A New Cognitive Architecture Based on Embodied Simulation
3.1 Theoretical Background on Embodied Cognition
3.1.1 The Convergence-Divergence Model
3.1.2 Embodied Concepts and Embodied Simulation
3.1.3 The Concept of Schema
3.1.4 Summary
3.2 Related Computational Models
3.2.1 Models Based on the Concept of Schema
3.2.2 Models of Embodied Simulation
3.3 A Cognitive Architecture Based on Embodied Simulation
3.3.1 The Schema System
3.3.2 The Motor-, Sensory- and Motivation Systems
3.3.3 Mechanics of the Building Blocks
3.3.4 Network Layout in the Schema System
3.4 Discussion
4 Integration of Internal Models by Making Use of Redundancies
4.1 Integration of Internal Models in Robotics
4.1.1 Approaches Based on Serialization
4.1.2 Approaches Based on Linear Combination
4.1.3 Approaches Based on Prioritization
4.2 Making Use of Redundancies for the Integration of Internal Models
4.2.1 Redundancy in Sensorimotor Tasks
4.2.2 Dynamic Selection of Solutions Using Dynamic Neural Fields
4.2.3 Distribibuted Decision Making in Co-ordinated DNFs
4.2.4 Summary
4.3 Using Networks of Sigma-Pi Units for the Learning and Query of Redundant Mappings, and for Robot Control
4.3.1 Networks of Sigma-Pi Units
4.3.2 Evaluation of the Sparsity in Networks of Sigma-Pi Units when Learning Kinematics Models
4.3.3 Using Multiple Queries for Distributed Decision Making
4.3.4 Using Networks of Sigma-Pi Units for Accurate Robot Control
4.4 Simulation Experiment with the iCub Humanoid Robot
4.5 Discussion
5 Self-Organized Learning of Multiple Internal Models
5.1 Bootstrapping the Learning of Internal Models by Exploiting Preliminary Model Predictions
5.2 Handling Noise
5.3 Example Application of Acquiring a Body-Schema
5.4 Discussion
6 Conclusion
6.1 Summary
6.2 Discussion in Relation to Machine Learning and the Field of Cognitive Architecture
6.2.1 Comparison with Other Cognitive Architectures
6.3 Discussion in Relation to Embodied Cognition and the Concept of Schema
6.4 Outlook
References
A Additional Mathematical Formulations