Motion Prediction for Human-Robot Collaborative Tasks Using LSTM
This study proposes an assistive robot system to reduce caregiving burdens in an aging society by supporting impaired body movements.The system focuses on bimanual tasks, such as pouring a drink from a bottle into a cup.Using 3D skeletal data excluding the impaired left hand, a deep learning model (LSTM) predicts the motion stages and 3D positions of the left hand, and the robot performs the substitute motions.The system uses data from multiple users to show its potential for improving patient independence and reducing caregiver workload.
Authors
- Toshihiro Nishimura (ORCID: https://orcid.org/0000-0003-2858-6368)
- Tokuo Tsuji (ORCID: https://orcid.org/0000-0002-1573-1928)
- Hiroaki Seki (ORCID: https://orcid.org/0000-0003-1588-4825)
- Yosuke Suzuki (ORCID: https://orcid.org/0000-0001-6862-9719)
- Tetsuyou Watanabe (ORCID: https://orcid.org/0000-0003-2549-1435)
- Tatsuhiro Hiramitsu (ORCID: https://orcid.org/0000-0002-1053-583X)
- K. Okada
Institutions
- Kanazawa University (JP)
Publication Details
- Journal
- Proceedings of International Conference on Artificial Life and Robotics
- Published
- 2025-02-13
- DOI
- https://doi.org/10.5954/icarob.2025.os15-4
- Primary Topic
- Human Pose and Action Recognition
- Type
- article
- Field-Weighted Citation Impact
- 0.00