Inhalt

•  
   1 Introduction
•  
   1.1 Objectives and contributions
•  
   1.2 Thesis outline
•  
  2 State of the art
•  
  2.1 Learning
•  
   2.1.1 Supervised
•  
   2.1.2 Unsupervised
•  
   2.1.3 Reward-based
•  
   2.2 Imitation
•  
  2.2.1 Biological background
•  
   2.2.1.1 Categorizing imitation
•  
   2.2.1.2 Imitation and memetics
•  
   2.2.1.3 Imitation in biology
•  
  2.2.2 Imitation in robotics
•  
   2.2.2.1 Challenges in robot imitation
•  
   2.2.2.2 Programming by demonstration
•  
   2.2.2.3 Imitation in multi-robot systems
•  
   2.2.3 Contrasting the thesis to the state of the art approaches
•  
   3 Architecture for learning and imitating in groups
•  
   3.1 Architectural overview
•  
   3.1.1 Motivation layer
•  
   3.1.2 Strategy layer
•  
   3.1.3 Skill layer
•  
   3.2 Layer interaction
•  
   3.3 Imitation in robot groups
•  
   3.4 Choice of the imitatee
•  
   3.5 Scenarios
•  
  4 Motivation layer
•  
   4.1 Background
•  
   4.1.1 Motivation in biological autonomous systems
•  
   4.1.2 Use of motivation in robots
•  
   4.2 Design of a robotic motivation system
•  
   4.2.1 Excitation
•  
   4.2.2 Prioritizing goals
•  
   4.3 Conclusion
•  
   5 Strategy layer
•  
  5.1 Background
•  
   5.1.1 Markov decision processes
•  
   5.1.1.1 Policy
•  
   5.1.1.2 Solving Markov decision processes
•  
   5.1.2 Semi-Markov decision processes
•  
  5.2 State of the art
•  
   5.2.1 Model-free approaches
•  
   5.2.2 Model-based approaches
•  
   5.2.3 Discussion
•  
   5.3 Policy
•  
   5.4 State abstraction
•  
  5.5 Model
•  
   5.5.1 Transition heuristic
•  
   5.5.2 Failure heuristic
•  
   5.5.3 Reward heuristic
•  
   5.5.4 Simplification heuristic
•  
   5.5.5 Experience heuristic
•  
   5.6 Sample frequency
•  
   5.7 Exploration
•  
   5.8 Example
•  
   6 Skill layer
•  
  6.1 Two modes of operation
•  
   6.1.1 Exploration mode
•  
   6.1.2 Exploitation mode
•  
   6.1.3 Interface with the environment
•  
   6.2 Component description
•  
   6.2.1 Skill manager
•  
   6.2.1.1 Skill generation
•  
   6.2.1.2 Skill ranking
•  
   6.2.1.3 Skill notification
•  
  6.2.2 Model manager
•  
   6.2.2.1 Creating and updating models
•  
   6.2.2.2 Scoring models
•  
   6.2.3 Error minimizer
•  
   6.3 Configuration
•  
   6.4 Conclusion
•  
  7 An integrative example
•  
   7.1 Implementation of the motivation layer
•  
   7.2 Implementation of the strategy layer
•  
   7.3 Implementation of the skill layer
•  
   7.4 Evaluation
•  
  8 Imitation in robot groups
•  
   8.1 Related work
•  
   8.2 Overview of the multi-robot imitation approach
•  
   8.3 Transforming observations
•  
   8.4 Understanding observed behavior
•  
   8.4.1 Viterbi
•  
   8.4.2 Interpreting observed behavior
•  
   8.4.3 Example
•  
   8.5 Integrating recognized behavior
•  
   8.6 Evaluation
•  
   8.6.1 CTF with three bases
•  
   8.6.2 CTF with five bases
•  
   8.7 Conclusion
•  
   9 Choice of the imitatee
•  
   9.1 Related work
•  
  9.2 Background
•  
   9.2.1 Bayesian networks and how to learn them
•  
   9.2.2 Affordances
•  
   9.3 Overview of the demonstrator choice process
•  
   9.4 Affordance detection
•  
   9.5 Affordance network generation
•  
  9.6 Comparing affordance networks
•  
   9.6.1 Structural difference of affordance networks
•  
   9.6.2 Parameter difference of affordance networks
•  
   9.6.3 Affordance network distance metric
•  
   9.7 Evaluation
•  
  9.7.1 Experimental setup
•  
   9.7.1.1 Parameterization of the environment
•  
   9.7.1.2 Affordances and their validation
•  
   9.7.1.3 Imitated behavior and how to measure its success
•  
  9.7.2 Selection experiment
•  
   9.7.2.1 Scenario
•  
   9.7.2.2 Procedure
•  
   9.7.2.3 Result
•  
  9.7.3 Robustness experiment
•  
   9.7.3.1 Scenario
•  
   9.7.3.2 Procedure
•  
   9.7.3.3 Result
•  
   9.7.4 Clustering experiment
•  
   9.7.4.1 Scenario
•  
   9.7.4.2 Procedure
•  
   9.7.4.3 Results
•  
   9.8 Conclusion
•  
   10 Summary and outlook
•  
   10.1 Summary
•  
   10.2 Contributions
•  
   10.3 Outlook
•  
   A Notation
•  
   B Algorithms
•  
   List of Figures
•  
   List of Tables
•  
   Own publications
•  
   Bibliography