著者

小畑 承経 熊谷 あやね 矢吹 佳子 東郷 俊太 姜 銀来 横井 浩史

出版者

一般社団法人 日本ロボット学会

雑誌

日本ロボット学会誌 (ISSN:02891824)

巻号頁・発行日

vol.39, no.8, pp.744-750, 2021 (Released:2021-10-23)

参考文献数

21

---

This study demonstrates that a bio-inspired fingertip structure improves the grasping performance of a prosthetic hand. A prosthetic hand attached to a human user requires a higher degree of grasping stability with minimal components compared to a typical robotic hand. Precision grasping tends to be less stable than power grasping; however, the former is used more frequently in daily life. To improve the stability of precision grasping, in this study, we focused on a two-layered elastic structure consisting of epidermis and subcutaneous tissue found in the finger. The efficiency of the two-layered structure, with respect to grasping stability, was evaluated by comparing one-layered and two-layered artificial fingertips. Finally, a two-layered elastic glove was designed based on the stability of the grasping. The stability of the precision grasping was evaluated using the pendulum experiment. The developed glove showed higher stability than the conventional prosthetic hand system.

著者

東郷 俊太 香川 高弘 宇野 洋二

出版者

一般社団法人電子情報通信学会

雑誌

電子情報通信学会技術研究報告 : 信学技報 (ISSN:09135685)

巻号頁・発行日

vol.112, no.480, pp.173-178, 2013-03-13

---

本研究では,人間が一次元の手先の目標軌道を追従する,一次元軌道追従タスクを行う動作を計測し,解析を行った.このタスクにおいて,追従方向と直交する方向はタスクの達成とは関係の無い方向である.計測結果として,終端姿勢において被験者の手先はタスク達成と関係の無い方向に大きくばらついた.また,接線速度の概形はベル型であり,小さなピークが多く含まれていた.UCM解析を用いて関節間協調を定量的に評価した結果,タスクの達成に影響を与えないばらつきが時間と共に増加した.これらの運動学的な特徴や関節間協調の特徴は,我々が提案したUCM参照フィードバック制御法によってよく再現された.UCM参照フィードバック制御法は,時々刻々と目標UCM空間を生成し,UCM空間上から最適な一点を選び続ける制御方策である.これらの結果は,人間はUCM空間を参照することによって,タスク達成に関係の無いばらつきを許容しながら,多関節を協調的に制御していることを示唆している.また,人間の視覚運動制御系においてUCM空間の神経表現が存在している可能性があることを示唆している.In this paper, we measured and analyzed a one-dimensional target tracking task in which human subjects tracked one-dimensional target trajectory of the hand. In this task, the direction orthogonal to the movement direction is task-irrelevant. As a result, the hand position at the movement end was more varied in the task-irrelevant direction. Moreover, the profile of tangential hand velocity was roughly bell-shaped but they also had small peaks. Results of UCM analysis which is used to quantify the joint coordination showed that the task-irrelevant variance increased with time. These characteristics of kinematics and joint coordination were well reproduced by our UCM reference feedback control. UCM reference feedback control generates a target UCM space each time step and selects one point in UCM space. These results suggest that the human refers to UCM space step by step to control a multi-joint arm allowing the task-irrelevant variability. Moreover, it is suggested that the human would represent UCM space in visuomotor control system.

著者

姜 銀来 李 文揚 陳 鵬 東郷 俊太 横井 浩史

出版者

一般社団法人 日本ロボット学会

雑誌

日本ロボット学会誌 (ISSN:02891824)

巻号頁・発行日

vol.38, no.7, pp.657-666, 2020 (Released:2020-09-05)

参考文献数

17

---

Humanoid robot arms have attracted more and more attention since it is human-friendly with the same size and movement characteristics of a human arm. A new 7-DoF (Degrees of Freedom) humanoid robot arm is proposed in this paper. To achieve high torques while keeping down the weight, all the motors are coupled with tendons. 2 motor 2-DoF (2M2D) and 3 motor 3-DoF (3M3D) coupled tendon-driven joint modules are proposed as basic components for the robot arm. The 2M2D coupled tendon-driven joint module and the 3M3D coupled tendon driven joint module structures are analyzed and compared. The 7 DoFs of a human arm were divided into a 3-DoF shoulder joint module, a 2-DoF elbow joint module, and a 2-DoF wrist joint module. The developed robot arm weighed 2.2[kg], while being able to lift a 1.5[kg] load. A current consumption experiment was conducted to verify the torque transmission characteristics of the 2M2D and 3M3D joint modules. A path repeatability experiment and an experiment to imitate dexterous manipulations of a human arm with master-slave control were conducted to investigate the performance of the modularized humanoid robot arm. The experimental results showed that the joint modules realized motor reallocation via tendon coupling, and that the humanoid robot arm was capable of performing dexterous manipulations like a human arm.