Estimation of Human Base Kinematics using Dynamical Inverse Kinematics and Contact-Aided Lie Group Kalman Filter

ICHR 2022 - Estimation of Human Base Kinematics

Reproducing the experiments

We provide a containerised virtual environment using Docker and Conda in order to launch the software in an isolated, reproducible manner. The dependencies and related versions used to build the environment can be checked in the deps folder. However, one can simply pull the pre-built docker image from the GitHub registry, since a docker.yaml in the .github/workflows folder dispatches a build workflow for the docker image.

1. Pull the docker image: bash docker pull ghcr.io/ami-iit/human-base-estimation-docker:latest

2. Launch the container: bash xhost + docker run -it --net=host --env="DISPLAY=$DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix" ghcr.io/ami-iit/human-base-estimation-docker:latest

Please be aware that xhost + is not a safe way to expose the X Server running on the host machine to the docker container. But in this scenario, we are not doing anything undesirable, so it is acceptable.

1. The experiment will start by automatically launching a yarp server, loading the dataset and launching the necessary applications along with the visualizer. The experiment will end and everything will close automatically.

For more details on the installation, implementation, and parameters configuration. please check KinDynFusion repository.

Known issues: Starting the docker daemon using Docker Desktop does not allow to display the visualizer GUI on the screen. I had to start the Docker daemon using dockerd with root privileges and then run the docker container also with root privileges in order to visualize the experiment.

Citing this work

@INPROCEEDINGS{10000199, author={Ramadoss, Prashanth and Rapetti, Lorenzo and Tirupachuri, Yeshasvi and Grieco, Riccardo and Milani, Gianluca and Valli, Enrico and Dafarra, Stefano and Traversaro, Silvio and Pucci, Daniele}, booktitle={2022 IEEE-RAS 21st International Conference on Humanoid Robots (Humanoids)}, title={Estimation of Human Base Kinematics using Dynamical Inverse Kinematics and Contact-Aided Lie Group Kalman Filter}, year={2022}, volume={}, number={}, pages={364-369}, doi={10.1109/Humanoids53995.2022.10000199}}

Maintainer

This repository is maintained by:

| | | | :----------------------------------------------------------: | :--------------------------------------------------: | | | @prashanthr05 |