Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bipedal Robot Running: Human-like Actuation Timing Using Fast and Slow Adaptations

Yusuke Sakurai, Tomoya Kamimura, Yuki Sakamoto, Shohei Nishii, Kodai Sato, Yuta Fujiwara, Akihito Sano

TL;DR

This work tackles the challenge of reproducing human-like muscle activation timing during running in a human-sized biped robot. It introduces a central pattern generator (CPG) with fast adaptation (phase resetting at foot touchdown) and slow adaptation (alignment of the estimated half-period $T_n^{e}$ with the actual half-period $T_n$), coupled with a pattern formulator that adjusts actuation timing via $\mu_N$ to stabilize thigh swing. Through both a simple SLIP model and robot experiments, the study shows that fast and slow adaptations yield phase-locked, human-like rhythmic activity, while actuation-timing adjustments extend sustained running and better match human joint timing. The findings support the view that human-like muscle activation timing emerges from adaptive CPG dynamics interacting with body-environment dynamics and suggest pathways to more robust, efficient bipedal robots and assistive devices.

Abstract

We have been developing human-sized biped robots based on passive dynamic mechanisms. In human locomotion, the muscles activate at the same rate relative to the gait cycle during running. To achieve adaptive running for robots, such characteristics should be reproduced to yield the desired effect, In this study, we designed a central pattern generator (CPG) involving fast and slow adaptation to achieve human-like running using a simple spring-mass model and our developed bipedal robot, which is equipped with actuators that imitate the human musculoskeletal system. Our results demonstrate that the CPG-based controller with fast and slow adaptations, and a adjustable actuator control timing can reproduce human-like running. The results suggest that the CPG contributes to the adjustment of the muscle activation timing in human running.

Bipedal Robot Running: Human-like Actuation Timing Using Fast and Slow Adaptations

TL;DR

This work tackles the challenge of reproducing human-like muscle activation timing during running in a human-sized biped robot. It introduces a central pattern generator (CPG) with fast adaptation (phase resetting at foot touchdown) and slow adaptation (alignment of the estimated half-period with the actual half-period ), coupled with a pattern formulator that adjusts actuation timing via to stabilize thigh swing. Through both a simple SLIP model and robot experiments, the study shows that fast and slow adaptations yield phase-locked, human-like rhythmic activity, while actuation-timing adjustments extend sustained running and better match human joint timing. The findings support the view that human-like muscle activation timing emerges from adaptive CPG dynamics interacting with body-environment dynamics and suggest pathways to more robust, efficient bipedal robots and assistive devices.

Abstract

We have been developing human-sized biped robots based on passive dynamic mechanisms. In human locomotion, the muscles activate at the same rate relative to the gait cycle during running. To achieve adaptive running for robots, such characteristics should be reproduced to yield the desired effect, In this study, we designed a central pattern generator (CPG) involving fast and slow adaptation to achieve human-like running using a simple spring-mass model and our developed bipedal robot, which is equipped with actuators that imitate the human musculoskeletal system. Our results demonstrate that the CPG-based controller with fast and slow adaptations, and a adjustable actuator control timing can reproduce human-like running. The results suggest that the CPG contributes to the adjustment of the muscle activation timing in human running.
Paper Structure (16 sections, 6 equations, 13 figures)

This paper contains 16 sections, 6 equations, 13 figures.

Table of Contents

  1. Introduction
  2. Method
  3. Bipedal robot based on passive dynamics
  4. Rhythm generator with fast and slow adaptations
  5. Pattern formulator for reproducing human-like actuation timing
  6. Simple model for verification of rhythm generator adaptation
  7. Results
  8. Effect of rhythm generator adaptation on simple model
  9. Robot experiments
  10. Fast and slow adaptation of rhythm generator
  11. Pattern formulator adjustment
  12. Discussion
  13. Effect of estimated gait cycle adaptation in rhythm generator
  14. Effect of actuation timing adjustment in pattern formulator
  15. Conclusion
  16. ...and 1 more sections

Figures (13)

  • Figure 1: (a) Bipedal running robot. BLDC motors in hip joints actuate thigh links. Muscle-tendon system represented by pneumatic actuators actuates knee and foot links. (b) Schematics of robot. Links are connected with active (solid red lines) or passive (dashed blue lines) wires and springs. (Colored version is available online.)
  • Figure 2: Schematics of controller. Left and right neural oscillators (CPGs) controls actuators according to their phases. Phases of CPGs are adjusted by sensory feedback and inter-limb interaction.
  • Figure 3: (a) Phases of rhythm generator are categorized into three phases: stance, early swing, and late swing. Pattern formulator switches control laws of actuators (hip joint motor and vastus pneumatic actuator) according to phase angle $\phi_i$ ($i = \mathrm{L, R}$). (b) Thigh angle $\theta$ is defined as angle between thigh link and vertical direction.
  • Figure 4: Simple SLIP model with leg dumper and hip actuator.
  • Figure 5: Time profiles of $y$ and $\dot{x}$ of SLIP model. Left and right figures indicate results for $(K_{\rm p},K_{\rm d})=(0,0)$ (without slow adaptation) and $(0.8,0.1)$ (with slow adaptation), respectively. Model falls down around 5 [s] without slow adaptation.
  • ...and 8 more figures