Characteristics, Management, and Utilization of Muscles in Musculoskeletal Humanoids: Empirical Study on Kengoro and Musashi

TL;DR

The paper addresses the lack of a unified framework for characterizing muscles in musculoskeletal humanoids. It introduces five properties (Redundancy, Independency, Anisotropy, Variable Moment Arm, Nonlinear Elasticity) and maps these to management/utilization strategies via reflex control and body-schema learning on Kengoro and Musashi. An integrated software-hardware system combining reflex controls, body-schema learning, muscle grouping, and adaptation is demonstrated through tasks including object grasping, rupture handling, cloth manipulation, and autonomous driving, highlighting potential for long-term stable, human-like movement. The work provides a foundational framework to guide design and control of future musculoskeletal humanoids and informs further hardware and cognitive development.

Abstract

Various musculoskeletal humanoids have been developed so far, and numerous studies on control mechanisms have been conducted to leverage the advantages of their biomimetic bodies. However, there has not been sufficient and unified discussion on the diverse properties inherent in these musculoskeletal structures, nor on how to manage and utilize them. Therefore, this study categorizes and analyzes the characteristics of muscles, as well as their management and utilization methods, based on the various research conducted on the musculoskeletal humanoids we have developed, Kengoro and Musashi. We classify the features of the musculoskeletal structure into five properties: Redundancy, Independency, Anisotropy, Variable Moment Arm, and Nonlinear Elasticity. We then organize the diverse advantages and disadvantages of musculoskeletal humanoids that arise from the combination of these properties. In particular, we discuss body schema learning and reflex control, along with muscle grouping and body schema adaptation. Also, we describe the implementation of movements through an integrated system and discuss future challenges and prospects.

Figures (9)

• Figure 1: The developed musculoskeletal humanoids: Kengoro and Musashi
• Figure 2: The basic musculoskeletal structure: the components include bones, joints, muscle actuators, wires, and nonlinear elastic elements. Muscle length $l$, muscle tension $f$, muscle temperature $c$, and joint angle $\theta$ (directly or indirectly depending on the robot) can be measured.
• Figure 3: The relationship between the characteristics of muscles and the advantages/disadvantages of the musculoskeletal structure.
• Figure 4: The overview of the system which manages and utilizes advantages/disadvantages of the musculoskeletal structure. The components include reflex control (Section \ref{['subsec:reflex']}), body schema learning (Section \ref{['subsec:learning']}), muscle grouping for body schema (Section \ref{['subsec:grouping']}), and body schema adaptation (Section \ref{['subsec:adaptation']}).
• Figure 5: The overview of the reflex control system. Muscle relaxation control, maximum speed control, muscle thermal control, antagonist inhibition control, and stretch reflex control alter the muscle length and the muscle stiffness in muscle stiffness control.