Braitenberg Vehicles as Developmental Neurosimulation

Area of Science:

• Neurobiology
• Artificial Intelligence
• Behavioral Science

Background:

• Traditional artificial and biological neural network models often treat the mature brain as a blank slate, neglecting developmental learning.
• Biological development and learning are crucial for understanding complex behaviors.

Purpose of the Study:

• To model the development of artificial organisms capable of exhibiting complex behaviors.
• To explore developmental embodied agents and their implementation.

Main Methods:

• Introduced three distinct approaches for developmental embodied agents: a Braitenberg vehicle-genetic algorithm hybrid, a multisensory Hebbian learning model, and multi-agent systems.
• Developed developmental Braitenberg vehicles (dBVs) to simulate emergent behaviors.

Main Results:

• The developed dBVs demonstrated a range of behaviors, including stimulus responses and collective motion.
• Showcased use cases like optimized spatial cognition, linking behavioral and neural models, and cumulative classification.

Conclusions:

• The developmental neurosimulation approach offers a novel perspective on artificial brain networks, embodied cognition, feedback, and emergence.
• Future applications in developmental neurosimulation are promising.