Predicting and attending to damaging collisions for placing everyday objects in photo-realistic simulations

DOI DOI Web Site 4 Citations 26 References
  • Aly Magassouba
    Advanced Speech Translation Research and Development Promotion Center, National Institute of Information and Communications Technology, Soraku, Kyoto, Japan
  • Angelica Nakayama
    Advanced Speech Translation Research and Development Promotion Center, National Institute of Information and Communications Technology, Soraku, Kyoto, Japan
  • Komei Sugiura
    Advanced Speech Translation Research and Development Promotion Center, National Institute of Information and Communications Technology, Soraku, Kyoto, Japan
  • Tsubasa Hirakawa
    Chubu Institute for Advanced Studies, Chubu University, Kasugai, Aichi, Japan
  • Hironobu Fujiyoshi
    Department of Robotic Science and Technology, College of Engineering, Chubu University, Kasugai, Aichi, Japan
  • Takayoshi Yamashita
    Department of Computer Science, College of Engineering, Chubu University, Kasugai, Aichi, Japan
  • Hisashi Kawai
    Advanced Speech Translation Research and Development Promotion Center, National Institute of Information and Communications Technology, Soraku, Kyoto, Japan

Bibliographic Information

Published
2021-04-16
Resource Type
journal article
DOI
  • 10.1080/01691864.2021.1913446
  • 10.48550/arxiv.2102.06507
Publisher
Informa UK Limited

Search this article

Description

Placing objects is a fundamental task for domestic service robots (DSRs). Thus, inferring the collision-risk before a placing motion is crucial for achieving the requested task. This problem is particularly challenging because it is necessary to predict what happens if an object is placed in a cluttered designated area. We show that a rule-based approach that uses plane detection, to detect free areas, performs poorly. To address this, we develop PonNet, which has multimodal attention branches and a self-attention mechanism to predict damaging collisions, based on RGBD images. Our method can visualize the risk of damaging collisions, which is convenient because it enables the user to understand the risk. For this purpose, we build and publish an original dataset that contains 12,000 photo-realistic images of specific placing areas, with daily life objects, in home environments. The experimental results show that our approach improves accuracy compared with the baseline methods.

18 pages, 7 figures, 5 tables. Submitted to Advanced Robotics

Journal

  • Advanced Robotics

    Advanced Robotics 35 (12), 787-799, 2021-04-16

    Informa UK Limited

Citations (4)*help

See more

References(26)*help

See more

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top