Ball Juggling on the Bipedal Robot Cassie

Abstract

The increasing integration of robots in daily life necessitates research in multitasking strategies. The act of juggling offers a simple platform to test techniques which may be generalizable to more complex tasks and systems. This paper presents both analytical and empirical results for successful ball-juggling on the bipedal robotic research platform Cassie. A control strategy inspired by mirror law algorithms was simulated on a simple paddle-ball system and then extended to the Cassie-ball system in simulations and experiments, using two low-level control schemes. A Poincaré analysis demonstrated́ stability for both controllers. Both simulated and experimental results show that the proposed strategy is robust to a wide range of physical parameters and that the act of juggling while balancing is achievable through multiple methods.