Soft-bodied Livings Simulation

Communication AI inspired by Jellyfish

Jellyfish are creatures that only show reflex reactions. We treat jellyfish as a system with high concentration, and use machine learning to emerge innovative communication ability. This soft embodiment artificial intelligence aims at establishing it as a powerful tool to give healing by using many information processing abilities possessed by a soft body.

Analysis of non-linear dynamics and chaos for Jellyfish model
applying vibration pattern by cellular automaton

Invertebrates are known to show autonomic behavior due to contraction of muscle cells even without the presence of a central control mechanism represented by the brain. The response to external stimuli of each muscle cell is a simple contraction, but the body structure becomes the constraint condition of the electric signal transmission path between nerve cells. Then, reactions are diffused throughout the cells with a time lag, creating various aspects of behavior. This study expresses the body structure of an organism by a non-linear spring-damper system using voxels as a vibration element, and handles a coupled oscillator system having soft physicality as an agent. We decide the vibration element group pattern based on CA and consider whether the behavior to be created is chaos (complex system). We also propose a phase model that responds to external stimuli and verify mechanism emergence.