Generating situation-dependent behavior: Decentralized control of multi-functional intestine-like robot that can transport and mix contents

Most robots are designed to perform a specific task in a predefined environment and have difficulty in producing situation-dependent behavior. To tackle this problem, we focus here on a mammal intestine that either transports or mixes the contents depending on the encountered circumstances. We propose a simple model for the intestinal movement and design an autonomous decentralized control scheme for an intestine-like robot by using coupled oscillators with local sensory feedback. Simulation results show that different types of motions are generated depending on the the physical conditions of the intestine and its contents. Our simulated robot does not require any input from a higher center to switch between different types of motions but determines autonomously which motion to generate. This study thus paves the way for developing "multi-functional robots" whose behavior is changed flexibly and spontaneously depending on circumstances.