Dyne

Code for our Humanoids '22 paper "Proprioceptive State Estimation with Kinematic Chain Modeling".

Repo Structure

• src: C++ source files including tests.
• python: Source files for building the python wrapper.
• tests: Python tests. Can evaluate pytest tests.
• scripts: Set of scripts to run various experiments.

Development

Dependencies

This project relies on

1. gtsam
2. gtdynamics

Please refer to their README for instructions on how to compile and build the libraries.

Compilation

We use CMake to configure the build directives, and make to build the library.

mkdir build && cd build
cmake ..
make -j8        # build
make -j8 check  # run tests

Python

Running CMake with the DYNE_BUILD_PYTHON_WRAPPER=ON flag will build the python wrapper.

This will create a new directory dyne in the root folder with the built python module. You can now run pip install -e . to install the python package.

State Estimators

We provide the following estimators:

1. BaseImu: This is an IMU-only dead-reckoning based estimator which relies on IMU preintegration. Liable to drift due to varying bias.
2. Bloesch: This is a factor-graph based implementation of the state estimator proposed by Bloesch et. al. in "Consistent Fusion of Leg Kinematics and IMU".
3. Dyne: Our state estimator proposed in the Humanoids 2022 paper.

Citing

If you use dyne in an academic context, please cite the following:

@INPROCEEDINGS{Agrawal22humanoids,
  author={Agrawal, Varun and Bertrand, Sylvain and Griffin, Robert and Dellaert, Frank},
  booktitle={2022 IEEE-RAS 21st International Conference on Humanoid Robots (Humanoids)}, 
  title={Proprioceptive State Estimation of Legged Robots with Kinematic Chain Modeling}, 
  year={2022},
  volume={},
  number={},
  pages={178-185},
  doi={10.1109/Humanoids53995.2022.10000099}}