Behavioral Repertoires for Soft Tensegrity Robots

TL;DR

This paper demonstrates a model-free, autonomous method for generating diverse locomotive behaviors in soft tensegrity robots, overcoming modeling challenges and expanding behavioral capabilities through real-world experimentation.

Contribution

It introduces a quality diversity algorithm that autonomously discovers a wide range of behaviors on physical soft robots without prior dynamics knowledge.

Findings

Generated diverse locomotive gaits useful for various tasks

Reduced reliance on simulation and manual tuning

Showed effectiveness of model-free, real-world optimization

Abstract

Mobile soft robots offer compelling applications in fields ranging from urban search and rescue to planetary exploration. A critical challenge of soft robotic control is that the nonlinear dynamics imposed by soft materials often result in complex behaviors that are counterintuitive and hard to model or predict. As a consequence, most behaviors for mobile soft robots are discovered through empirical trial and error and hand-tuning. A second challenge is that soft materials are difficult to simulate with high fidelity -- leading to a significant reality gap when trying to discover or optimize new behaviors. In this work we employ a Quality Diversity Algorithm running model-free on a physical soft tensegrity robot that autonomously generates a behavioral repertoire with no a priori knowledge of the robot dynamics, and minimal human intervention. The resulting behavior repertoire displays a diversity of unique locomotive gaits useful for a variety of tasks. These results help provide a road map for increasing the behavioral capabilities of mobile soft robots through real-world automation.