Shock mitigation and recovery motion for a falling small humanoid robots

Some robots such as humanoids are very unstable and have a high risk of fall over and hardware damage. Therefore, we have been working on various mechanisms to mitigate the impact of fall over by attaching to a small humanoid robot.

Shock minitagation by utillizing umbrella

When the humanoid robot falls, it opens a folding umbrella attached to its head to support itself and mitigate the impact. The robot detects and predicts the fall by a sensor, and opens the umbrella to mitigate the impact of fall over. After that, the robot selects a recovery action corresponding to own state, and finally closes the umbrella to return to the standing state. By utilizing a folding umbrella, we have been realized a reversible shock mitigation mechanism.

Shock minitagation by utillizing soft wheel

This study have been working on a shock mitigation mechanism with soft wheels. the robot can mitigate the impact of a fall without dropping an object grasped with the hands by attaching soft wheels to the shoulders. When the robot falls over on uneven terrain, it rises up after moving to level ground on its wheels. In this way, the robot can avoid another fall on uneven terrain by passing such terrain with wheels.


Videos

Shock minitagation by utillizing umbrella

Recovery corresponding to the states of fall over

Closing the umbrella after recovery

Recovery corresponding to fall over on walking

Shock minitagation by utillizing soft wheel

Recovery from fall over in a step

Recovery from fall over on a slope


Research poster (in Japanese)

  1. Taiki Masubuchi, Shunsuke Kudoh and Kohei Kimura. Shock Mitigation and Fall Recovery for Humanoid Robot Utilizing Opening and Closing Mechanism of a Folding Umbrella. Transactions of the Society of Instrument and Control Engineers. Vol.61 No.1 pp.32-45

  2. Taiki Masubuchi, Shunsuke Kudoh, Kohei Kimura. Shock mitigation and recovery motion for a falling robot utilizing a folding umbrella. 24th SICE System Integration Division Annual Conference (SI2023). 2023. 3H2-03