Related Experiment Videos
Modelling brain emergent behaviours through coevolution of neural agents
Michail Maniadakis1, Panos Trahanias
1Institute of Computer Science, Foundation for Research and Technology-Hellas (FORTH), P.O. Box 1385, Heraklion, 711 10 Crete, Greece. mmaniada@ics.forth.gr
Summary
This study presents a novel computational framework for integrating partial brain models in artificial organisms. The agent-based approach enhances robotic behavioral capabilities through improved substructure integration.
Area of Science:
- Computational Neuroscience
- Artificial Intelligence
- Robotics
Background:
- Existing artificial brain models face integration challenges due to heterogeneity.
- Developing integrated computational models is crucial for advancing artificial organism cognition.
Purpose of the Study:
- To introduce a computational framework for brain modeling that emphasizes integrative performance of substructures.
- To embed these models in a robotic platform to enhance behavioral capabilities.
Main Methods:
- An agent-based approach was used for designing autonomous substructures.
- A collaborative coevolutionary algorithm was employed for specifying agent structures.
- Computational models for the motor cortex and hippocampus were implemented.
Main Results:
- The framework successfully integrated partial brain models.
- Implemented models were embedded in a simulated mobile robot.
- The approach demonstrated enhanced behavioral capabilities in the robot.
Conclusions:
- The proposed framework effectively addresses the integration of heterogeneous brain substructures.
- Agent-based modeling and coevolutionary algorithms facilitate the development of autonomous and cooperative brain systems.
- The successful implementation on a robotic platform validates the approach for supporting artificial organism cognition.