ILC-driven control enhancement for integrated MIMO soft robotic system

Abstract

This study presents a methodology employing Iterative Learning Control (ILC) to enhance the control performance of soft grippers equipped with multiple curvatures and variable stiffness. ILC is a learning-based control approach that progressively reduces errors in repetitive tasks, known for delivering superior performance in complex systems. In the context of the increasing utilization of robotic technology across various industries, the control technology of soft robots, especially soft grippers with multiple curvatures and variable stiffness, is a crucial issue. While prior research has focused on single-curvature and single-input single-output (SISO) systems, this study addresses the intricate control problem of multi-input multi-output (MIMO) soft gripper systems capable of multiple curvatures. It also proposes an enhanced design for soft grippers with multiple curvatures and variable stiffness while highlighting the potential of ILC for enhancing control performance.