Pose Estimation of a Cable-Driven Serpentine Manipulator Utilizing Intrinsic Dynamics via Physical Reservoir Computing

IROS2025

TL;DR explore a new approach to pose estimation, where intrinsic nonlinear dynamics of a lightweight cable-driven serpentine manipulator are exploited via physical reservoir computing.

Overview

We propose a pose estimation method for a lightweight 9-DOF cable-driven serpentine manipulator, utilizing physical reservoir computing (PRC) to harness the manipulator's intrinsic nonlinear dynamics. By placing all motors and sensors at the base, we reduce the moving weight to only 308 g. Our PRC-based method achieves a mean pose error of 4.3 mm, comparable to an LSTM baseline (4.4 mm), and significantly better than analytical modeling (39.5 mm). This approach opens new directions for control and perception in lightweight, flexible cable-driven robots.

Citation

@inproceedings{tanaka2025pose,
  title={Pose Estimation of a Cable-Driven Serpentine Manipulator Utilizing Intrinsic Dynamics via Physical Reservoir Computing},
  author={Tanaka, Kazutoshi and Takahashi, Tomoya and Hamaya, Masashi},
  booktitle={2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2025},
  organization={IEEE}
}