Utilizing joint redundancy and muscle redundancy for motion planning of musculoskeletal manipulator

The human musculoskeletal structure has two types of redundancies: joint redundancy and muscle redundancy. Joint redundancy allows appropriate posture adjustments for situational demands while performing tasks, whereas muscle redundancy facilitates the balance of tension and relaxation between opposing muscles to modify whole-body flexibility. Humans utilize these redundancies to perform tasks involving interactions with the environment. Musculoskeletal robots are expected to be used in living spaces through the application of such musculoskeletal characteristics. However, engineering approaches that utilize these redundancies remain unexplored. This is because considering both redundancies simultaneously is equivalent to solving a nested complex optimization problem. This study proposes an optimization framework that integrates joint and muscle redundancies by optimizing robot posture with consideration of muscle balance and performance. In a case study, we identify postures that satisfy reference muscle stiffness and confirm adaptability to changes in stiffness levels. The contribution of this study is to enable an expanded search that includes posture adjustments unattainable with conventional sequential approaches, thereby providing a new guideline for posture optimization in musculoskeletal robotics.