Growing and Evolving Vibrationally Actuated Soft Robots

Abstract

Designing soft robots is difficult, time-consuming, and non-intuitive. Soft robot design faces two main challenges: structure and control. This research uses generative encodings to grow structures and a vibrational mechanism to control locomotion. In this paper, we demonstrate the ability to successfully evolve soft robots that can move when vibrated. Soft bodies are grown through a grammatical process and simulated in the Bullet physics engine. We also briefly outline a method of evolving scalable solutions that we are currently investigating. It should be capable of generating soft robots of various sizes that can move when vibrated.