Explore the Behavioral Repertoire of a Wireless Tensegrity robot by MAPElites Algorithms

Tensegrity robots are biology inspired soft-robots that constructed by rigid (struts) elements and tensile (cables and springs) elements. It's structure design provides great strength-to-weight ratio and could navigate and function under complex environments. However, the wonderful design of the tensegrity structure also brings complex dynamics to the robot which makes the robot hard to control. Here at Union, we built a Wireless Vibrationally Actuated Limbless Tensegrity Robot (Valtr) for our experiment . In order to learn the methods to control the robot, we first look at the different behaviors that it could perform with given commands (frequency combinations). In my presentation, I will talk about how we use a Quality Diversity (QD) algorithm to explore the behavior repertoire (translation and rotation) of our tensegrity robot Valtr.