Acquisition of Synergy for Low-dimensional Control of Multi-fingered Hands by Reinforcement Learning

DOI

説明

Synergy is the method that reduces the control inputs of a multi-fingered hand and is utilized for designing underactuated robotic hands and efficient control. Calculating conventional synergies depends on the measured human grasping postures. Therefore, preparing synergies for the not-human-like multi-fingered hands is challenging. We propose a reinforcement learning platform for acquiring synergies of a multi-fingered robotic hand through learning a grasping task. The learning process automatically generates postures for creating synergies so that this system can prepare synergies for any robotic hand. Experiments show that this reinforcement learning platform improves learning tasks and acquires the synergy that is suitable for the learned task.

収録刊行物

キーワード

詳細情報 詳細情報について

  • CRID
    1390296808234315264
  • DOI
    10.5954/icarob.2023.os16-1
  • ISSN
    21887829
  • 本文言語コード
    en
  • データソース種別
    • JaLC
    • Crossref
    • OpenAIRE
  • 抄録ライセンスフラグ
    使用不可

問題の指摘

ページトップへ