These notes are a summary of concepts presented in “AIBO: Toward the Era of Digital Creatures.”
Fujita, M. (2000). AIBO: Towards the Era of Digital Creatures. In: Hollerbach, J.M., Koditschek, D.E. (eds) Robotics Research. Springer, London. https://doi.org/10.1007/978-1-4471-0765-1_38
- Design and Evaluation Considerations
- Maximize lifelike appearance through behavior and motion
- Semantic Differential (SD) for final product assessment
- Computation time and feedback latency as measures for behavior types
- Reflexive behaviors emphasize speed, while deliberate behaviors focus on careful planning
- Behavioral Complexity
- Increase in behavior diversity ensures nonrepetition in similar situations
- Introduction of artificial instincts and emotions to drive unique behaviors
- Factors contributing to complexity
- Multiple motivations for movement
- High degrees of freedom in motion
- Nonrepeated behavioral exhibition
- Behavior Generation Framework
- Behavior generation processes
- Fusion of reflexive and deliberate behaviors across different timescales
- Integration of internal motivations with external stimuli
- Coordination of instincts and emotions with observed stimuli
- Behavior generation processes
- Internal Status and External Stimuli
- Internal states (instincts/emotions) influence responses to external stimuli
- Internal states are dynamically updated based on stimuli and interactions
- Adaptation and Learning
- Long-term adaptation through learning to enhance behavioral complexity
- Observation over time reveals evolving, nonrepetitive behaviors
- Behavioral Classification
- Reflexive behaviors
- Quick, stimulus-driven responses with minimal computation time
- Use simple, noniterative decision rules
- Deliberate behaviors
- Slower, planned actions requiring database searches or inference
- Longer computational time for decision-making
- Reflexive behaviors
- Behavioral Layers
- Motor command layer
- Generates reflexive behaviors using sensor feedback and commands
- Action sequence layer
- Resolves mechanical constraints and posture transitions
- Behavior generation layer
- Combines external events with internal states to create behaviors
- Motor command layer
- Emotion and Sensory Processing
- Evaluating emotions using sensor inputs
- Assigning dynamics to emotions for behavior modeling
- Modules for perception include
- Vision (color and obstacle detection)
- Sound processing
- Posture analysis
- Agent Architecture
- Behavior-based architecture featuring
- Randomness
- Instincts and emotions
- Learning abilities
- Development capabilities
- Diverse motion generation
- Behavior-based architecture featuring