Simple decentralized control mechanism that enables limb adjustment for adaptive quadruped running

Akira Fukuhara, Yukihiro Koizumi, Tomoyuki Baba, Shura Suzuki, Takeshi Kano, Akio Ishiguro

Research output: Contribution to journalArticlepeer-review

Abstract

Quadrupeds exhibit versatile and adaptive running by exploiting the flying phase during the stride cycle. Various interlimb coordination mechanisms focusing on mechanical loads during the stance phase have been proposed to understand the underlying control mechanism, and various gait patterns have been reproduced. However, the essential control mechanism required to achieve both steady running patterns and non-steady behaviours, such as jumping and landing, remains unclear. Therefore, we focus on the vertical motions of the body parts and propose a new decentralized interlimb coordination mechanism. The simulation results demonstrate that the robot can generate efficient and various running patterns in response to the morphology of the body. Furthermore, the proposed model allows the robot to smoothly change its behaviour between steady running and non-steady landing depending on the situation. These results suggest that the steady and non-steady behaviours in quadruped adaptive running may share a common simple control mechanism based on the mechanical loads and vertical velocities of the body parts.

Original languageEnglish
Article number20211622
JournalProceedings of the Royal Society B: Biological Sciences
Volume288
Issue number1962
DOIs
Publication statusPublished - 2021

Keywords

  • Central pattern generator
  • interlimb coordination
  • local sensory feedback
  • quadruped running

ASJC Scopus subject areas

  • Immunology and Microbiology(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Environmental Science(all)
  • Agricultural and Biological Sciences(all)

Fingerprint

Dive into the research topics of 'Simple decentralized control mechanism that enables limb adjustment for adaptive quadruped running'. Together they form a unique fingerprint.

Cite this