Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hardware Design and Learning-Based Software Architecture of Musculoskeletal Wheeled Robot Musashi-W for Real-World Applications

Kento Kawaharazuka, Akihiro Miki, Masahiro Bando, Temma Suzuki, Yoshimoto Ribayashi, Yasunori Toshimitsu, Yuya Nagamatsu, Kei Okada, and Masayuki Inaba

TL;DR

This work tackles the challenge of deploying musculoskeletal humanoids to real-world tasks by combining a musculoskeletal dual-arm design with a mechanum-wheeled base and a learning-based software stack. The hardware integrates flexible upper limbs with configurable muscle arrangements, a linear motion mechanism, and a sensor-rich torso, while the software combines static and dynamic body schemas, reflex controls, and visual feedback to plan and execute actions. Key contributions include a hardware-software integration framework, online adaptation through static and dynamic body schemas, and reflex-based safety that manages internal forces and motor temperature; the authors demonstrate cleaning via human teaching, carrying heavy objects with muscle addition, and dynamic cloth manipulation to set a table. The results indicate that exploiting muscle redundancy, variable stiffness, and modular attachments enables practical manipulation tasks, suggesting a viable path toward real-world musculoskeletal robotics.

Abstract

Various musculoskeletal humanoids have been developed so far. While these humanoids have the advantage of their flexible and redundant bodies that mimic the human body, they are still far from being applied to real-world tasks. One of the reasons for this is the difficulty of bipedal walking in a flexible body. Thus, we developed a musculoskeletal wheeled robot, Musashi-W, by combining a wheeled base and musculoskeletal upper limbs for real-world applications. Also, we constructed its software system by combining static and dynamic body schema learning, reflex control, and visual recognition. We show that the hardware and software of Musashi-W can make the most of the advantages of the musculoskeletal upper limbs, through several tasks of cleaning by human teaching, carrying a heavy object considering muscle addition, and setting a table through dynamic cloth manipulation with variable stiffness.

Hardware Design and Learning-Based Software Architecture of Musculoskeletal Wheeled Robot Musashi-W for Real-World Applications

TL;DR

This work tackles the challenge of deploying musculoskeletal humanoids to real-world tasks by combining a musculoskeletal dual-arm design with a mechanum-wheeled base and a learning-based software stack. The hardware integrates flexible upper limbs with configurable muscle arrangements, a linear motion mechanism, and a sensor-rich torso, while the software combines static and dynamic body schemas, reflex controls, and visual feedback to plan and execute actions. Key contributions include a hardware-software integration framework, online adaptation through static and dynamic body schemas, and reflex-based safety that manages internal forces and motor temperature; the authors demonstrate cleaning via human teaching, carrying heavy objects with muscle addition, and dynamic cloth manipulation to set a table. The results indicate that exploiting muscle redundancy, variable stiffness, and modular attachments enables practical manipulation tasks, suggesting a viable path toward real-world musculoskeletal robotics.

Abstract

Various musculoskeletal humanoids have been developed so far. While these humanoids have the advantage of their flexible and redundant bodies that mimic the human body, they are still far from being applied to real-world tasks. One of the reasons for this is the difficulty of bipedal walking in a flexible body. Thus, we developed a musculoskeletal wheeled robot, Musashi-W, by combining a wheeled base and musculoskeletal upper limbs for real-world applications. Also, we constructed its software system by combining static and dynamic body schema learning, reflex control, and visual recognition. We show that the hardware and software of Musashi-W can make the most of the advantages of the musculoskeletal upper limbs, through several tasks of cleaning by human teaching, carrying a heavy object considering muscle addition, and setting a table through dynamic cloth manipulation with variable stiffness.
Paper Structure (15 sections, 6 equations, 9 figures)

This paper contains 15 sections, 6 equations, 9 figures.

Table of Contents

  1. INTRODUCTION
  2. Hardware of Musculoskeletal Wheeled Robot Musashi-W
  3. Design of the Musculoskeletal Dual Arms
  4. Design of Wheeled Base
  5. Design of Musculoskeletal Wheeled Robot Musashi-W
  6. Software of Musculoskeletal Wheeled Robot Musashi-W
  7. Static Body Schema Learning
  8. Dynamic Body Schema Learning
  9. Reflex Control
  10. Visual Recognition and Classical Control
  11. Experiments
  12. Duster Experiment with Human Teaching
  13. Object Carrying Experiment Considering Muscle Addition
  14. Table Setting Experiment with Dynamic Cloth Manipulation
  15. CONCLUSION

Figures (9)

  • Figure 1: The newly developed musculoskeletal wheeled robot Musashi-W.
  • Figure 2: The hardware overview of Musashi-W.
  • Figure 3: The circuit configuration of Musashi-W.
  • Figure 4: The software overview of Musashi-W.
  • Figure 5: Parameter configuration of the wheeled base with four mechanum wheels.
  • ...and 4 more figures