日本ロボット学会誌
Online ISSN : 1884-7145
Print ISSN : 0289-1824
ISSN-L : 0289-1824
39 巻, 3 号
選択された号の論文の28件中1~28を表示しています
特集について
展望
解説
国際会議報告
論文
  • 森安 竜大, 橋本 俊哉, 日下部 信一, 加嶋 健司
    2021 年 39 巻 3 号 p. 259-262
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー

    Dynamical system learning, also known as system identification, is a powerful tool for modeling unknown systems from data. However, guaranteeing properties such as stability of the obtained model is generally difficult, especially for complex nonlinear models. Therefore, for industrial applications, a reliable learning method that can ensure that the model has some desirable properties is required. In this paper, a novel method for learning a nonlinear model with guaranteed topological properties is proposed. The method employs a model structure represented by transforming the internal state, defined by an internal model with simple known dynamic properties, by a homeomorphic map. This enables to learn a model that guarantees not only the stability of the model but also the existence of limit cycles and other topological properties of the state trajectories in state space.

  • 丸山 孝仁, 小川 翔太, 武田 信和
    2021 年 39 巻 3 号 p. 263-266
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー

    Providing three-dimensional information to operators is a vital technology to operate robotic systems remotely. ITER, which is the world's largest fusion energy experiment, requires maintenance by robotics devices operated remotely because of its radiation environment. Applicable sensors are highly restricted due to the radiation. Authors propose a method using one moving camera and multi-view stereo technology to obtain point clouds as three-dimensional information for operators. As multi-view stereo technologies, Structure-from-Motion and Patch-Based Multi-view Stereo are adopted. Authors tested the proposed method and confirmed the method is capable to create point clouds successfully.

  • 本蔵 紅奈, 高橋 正樹
    2021 年 39 巻 3 号 p. 267-270
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー
    電子付録

    We propose shared control that enables complicated and diverse motions. It is essential to support human operations by considering human decision making and motion constraints to realize complex motions. Furthermore, when various movements are targeted, it is necessary to judge the movement that the human wants. In this study, we assess the desired movement based on the affordances defined as probabilities. Furthermore, the reference motions are learned stochastically with Probabilistic Movement Primitives (ProMPs). It assists human operations by considering the human decision making and the motion constraints by the variances of the references. The proposed method was evaluated through experiments.

  • 岸波 華彦, 糸山 克寿, 西田 健次, 中臺 一博
    2021 年 39 巻 3 号 p. 271-274
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー

    In recent years, many kinds of sensors have been studied to recognize the environment, and they are used for AR and VR applications and for SLAM. Although ultrasonic signals with high directivity and high resolution are often used, there are problems such as ultrasonic exposure and grating noise at the rising edge. In this paper, we propose a new active sensing method based on audible sounds that is robust to environmental noise by combining weighting likelihood functions and standing waves. Compared to ultrasonic signals, audible sound tends to spread out, which leads to misalignment of distance estimates and loss of map consistency over time. Therefore, we derive the effective azimuth angle based on the directional characteristics of the speaker and calculate the likelihood of the presence or absence of obstacles using the observation model. In addition, we introduce occupancy grid mapping to produce a map that best explains the estimated distances. We performed real-world two-dimensional environment recognition experiments using the proposed method to detect and map surrounding obstacles, and showed the effectiveness of the method.

  • 八木 秀明, 星野 智史
    2021 年 39 巻 3 号 p. 275-278
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー

    For autonomous mobile robot navigation, localization is a fundamental technique. For this purpose, environmental maps are usually given beforehand by using SLAM. In stead of SLAM, we use an existing geographical map. Since road and building information are already included in the map, we focus on the scan data of the actual road surface and buildings in the environment obtained from a 3D LiDAR. In the map, MCL technique enables a robot to estimate the position. In the experiments, we show that the robot is able to autonomously move more stably through the localization with not only the scan data of the road surface, but also the ones of the buildings.

  • 清川 拓哉, 佐久間 達也, 高松 淳, 小笠原 司
    2021 年 39 巻 3 号 p. 279-282
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー

    To achieve a general-purpose assembly robot system that can handle various shaped objects, we propose a soft jig that deforms according to the shape and the behavior of assembly parts. The soft jig is based on a jamming gripper used for robot manipulation as a general-purpose robotic gripper developed in the field of soft robotics. The usability of the soft jig was evaluated from the viewpoint of versatility for fixing parts in assembly operations.

  • —視覚的力触覚を適用した操作性の脳波計測評価—
    春名 正樹, 川口 昇, 勝又 徳也, 荻野 正樹, 小池 俊昭
    2021 年 39 巻 3 号 p. 283-286
    発行日: 2021年
    公開日: 2021/04/28
    ジャーナル フリー
    電子付録

    We previously proposed a visual haptics technology which overlays visual images on the contact points of the remote machine hand with the augmented reality (AR) technology and demonstrated the effectiveness through in-house virtual reality (VR) experiments across multiple subjects. In this paper, we introduce our additional VR training/verification application of the remote machine system for an effective design and operation, to show more rigorous evaluation results of training effect by the proposed visual haptics through the analysis of brain wave signals. It is experimentally verified that our method can significantly reduce information flow in the brain analyzed by Smooth Coherence Transform (SCoT) and Cognitive Load Index (CLI) up to 34% and 43%, respectively. It suggests that the proposed visual haptics can contribute to improving the performance of our remote machine system by decreasing the task burden of user's manipulation.

feedback
Top