FURI | Spring 2025

Trajectory Generation and Design of Experiment for Training a Physical Reservoir Computer

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

Soft actuators and reconfigurable robots have gained recognition for their adaptability and safety in human interaction. However, their inherent softness makes precise control challenging and traditional training of neural networks often falls short. This research aims to design an experiment which focuses on generating a trajectory for the experimental soft actuator to study its behaviors for physical reservoir computing. This study will create a pathway of study for other soft robots to enhance their sensing capabilities and control.