Autonomous Underwater Manipulation Based on the Underwater Bilateral Control and Machine Learning

项目来源

日本学术振兴会基金(JSPS)

项目主持人

元井 直樹

项目受资助机构

神戸大学

立项年度

2024

立项时间

未公开

项目编号

24K00869

研究期限

未知 / 未知

项目级别

国家级

受资助金额

18590000.00日元

学科

電力工学関連

学科代码

未公开

基金类别

基盤研究(B)

关键词

モーションコントロール ; 制御工学 ; ロボティクス ; 水中ロボット ; 機械学習 ;

参与者

未公开

参与机构

未公开

项目标书摘要:Outline of Research at the Start:本研究では水中ロボットにおける水中マニピュレーションの自動化を目指す。まず、高精度な力覚伝送を実現する水中バイラテラル制御を開発し、水中マニピュレーションにおける直感的な遠隔操作を実現する。また、水中バイラテラル制御における操作者の遠隔操作情報(位置・力情報)および視覚情報を抽出する。抽出した人間の遠隔操作情報および視覚情報ををもとにモーションコピー技術に機械学習を融合することで、人間操作と同程度の性能を有する自律型水中マニピュレーション技術を実証する。

  • 排序方式:
  • 1
  • /
  • 1.Research on Extraction and Reproduction of Human Calligraphy Motion Using Motion Sensor and 6-axis Force Sensor

    • 关键词:
    • Copying;Industrial robots;6-axis force sensor;Extraction systems;Force data;Human motions;Industrial fields;Motion sensors;Parallel robots;Position data;Robots system;System use
    • Asada, Ryo;Motoi, Naoki
    • 《2025 IEEE International Conference on Mechatronics, ICM 2025》
    • 2025年
    • February 28, 2025 - March 2, 2025
    • Wollongong, NSW, Australia
    • 会议

    In recent years, robot systems to reproduce human motions such as a motion-copying system have been studied in the industrial field. The motion-copying system uses bilateral control to extract human motions. However, the robot's mechanics may affect human operability. This paper proposes the human motion extraction system using a motion sensor and a force sensor. The proposed system does not interfere with human operation. This study focuses on human calligraphy motions as the dexterous human motions. The extraction system extracts the brush position data from a motion sensor. In addition, the system extracts the force data applied to the brush from a force sensor. By using the extracted position and force data, a parallel robot reproduces human calligraphy motions. The reproducibility of the skills and the effectiveness of this system were confirmed by comparing the original characters with the reproduced characters. © 2025 IEEE.

    ...
  • 2.Research on Estimation Method for Dynamic Characteristics of Micro Objects Using Micro-Macro Bilateral Control

    • 关键词:
    • Micromanipulators;Motion control;Robotics;Viscosity;Bilateral control;Dynamics characteristic;Estimation methods;Measurements of;Micro manipulation;Micro-macro;Micro-objects;Parameters estimation;Property;Recursive least squares
    • Arikawa, Shohei;Motoi, Naoki
    • 《34th IEEE International Symposium on Industrial Electronics, ISIE 2025》
    • 2025年
    • June 20, 2025 - June 23, 2025
    • Toronto, ON, Canada
    • 会议

    In recent years, measurement of physical properties such as dynamic characteristics of micro objects is required in industrial and medical fields. However, measurement of dynamic characteristics such as stiffness, viscosity, and mass for micro objects has not been widely studied. Moreover, contacting the objects by manipulators was performed without force feedback in the conventional methods. Therefore, there is a possibility of damaging the micro objects. To address these problems, this paper proposes an estimation method for the dynamic characteristics of the micro objects using micro-macro bilateral control. The micro-macro bilateral system allows the operator to perform the micromanipulation of the objects safely. From position, velocity, acceleration, and force information obtained by grasping, the dynamic characteristics are estimated by the recursive least squares (RLS) algorithm. As a result, the characteristics such as the stiffness, the viscosity, and the mass of the micro objects are obtained based on the micro-macro bilateral control. The validity of the proposed method was confirmed by the experimental results. © 2025 IEEE.

    ...
  • 排序方式:
  • 1
  • /