tensegrity_perception

This is the official repository for the paper "6N-DoF Pose Tracking for Tensegrity Robot". Paper Link

This paper has been presented at the International Symposium on Robotics Research. (ISRR 2022)

Introduction

This work aims to address the pose tracking problem of an N-bar tensegrity robots through a markerless, vision-based method, as well as novel, on-board sensors that can measure the length of the robot's cables. In particular, an iterative optimization process is proposed to estimate the 6-DoF poses of each rigid element of a tensegrity robot from an RGB-D video as well as endcap distance measurements from the cable sensors. To ensure the pose estimates of rigid elements are physically feasible, i.e., they are not resulting in collisions between rods or with the environment, physical constraints are introduced during the optimization.

Installation

Install Miniconda/Anaconda from here
conda env create -f environment.yml
conda activate tensegrity

Demo Trajactories

Some demo trajactories could be downloaded here.

Usage

Create a dataset folder in the repository and place the demo trajectories inside.
A bash script tracking.sh is provided, which contains commands to run the algorithm, evaluate, and generate videos for visualization.

(Deprecated)

The original implementation (which requires CUDA and is much slower) accompanying the ISRR submission is in the ISRR_version branch.

Citation

@misc{lu20226ndof,
    title={6N-DoF Pose Tracking for Tensegrity Robots},
    author={Shiyang Lu and William R. Johnson III and Kun Wang and Xiaonan Huang and Joran Booth and Rebecca Kramer-Bottiglio and Kostas Bekris},
    year={2022},
    eprint={2205.14764},
    archivePrefix={arXiv},
    primaryClass={cs.RO}
}