From slime molds to soft deformable robots

An autonomous decentralized control mechanism, where the coordination of simple individual components yields non-trivial macroscopic behavior or functionalities, is a key to understanding how animals orchestrate the large degrees of freedom of their bodies in response to different situations. However, a systematic design methodology is still missing. To alleviate this problem, we focus, in this chapter, on the plasmodium of a true slime mold (Physarum polycephalum), which is a primitive multinucleate single-cell organism. Despite its primitiveness, and lacking a brain and nervous system, the plasmodium exhibits surprisingly adaptive and versatile behavior (e.g. taxis, exploration). This ability has undoubtedly been honed by evolutionary selection pressure, and there likely exists an ingenious mechanism that underlies the animals’ adaptive behavior. We successfully extracted a design scheme for decentralized control and implemented it in an amoeboid robot with many degrees of freedom. The experimental results showed that adaptive behaviors emerge even in the absence of any centralized control architecture.