Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/agimus-project/video_guided_tamp_planner

Planner for the ICRA 2023 submission titled Multi-Contact Task and Motion Planning Guided by Video Demonstration.

https://github.com/agimus-project/video_guided_tamp_planner

Science Score: 13.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (9.0%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Planner for the ICRA 2023 submission titled Multi-Contact Task and Motion Planning Guided by Video Demonstration.

Basic Info
  • Host: GitHub
  • Owner: agimus-project
  • License: bsd-3-clause
  • Language: Python
  • Default Branch: main
  • Homepage:
  • Size: 116 KB
Statistics
  • Stars: 4
  • Watchers: 3
  • Forks: 1
  • Open Issues: 0
  • Releases: 0
Created almost 3 years ago · Last pushed over 1 year ago
Metadata Files
Readme License

README.md

Multi-Contact Task and Motion Planning Guided by Video Demonstration

Installation

Please follow the installation procedure of GuidedTAMPBenchmark repo. bash conda config --add channels conda-forge conda activate gtamp # activate your GuidedTAMPBenchmark environment conda install hpp-gepetto-viewer -y # installs HPP libraries conda install pycollada anytree quaternion # additional planner dependencies

Data preprocessing

To run the planning there are two possibilities: 1. Run the algorithm on data files from GuidedTAMPBenchmark repo (no data preprocessing needed) 2. Replace the GuidedTAMPBenchmark data files with your custom data

The steps to create the data files are the following: - shoot the video where human hands and objects are clearly visible at times of contact - process the video with 6D pose estimator (f.e. CosyPose) - process the video with hand-object contact recognizer (f.e. Hand Object Detector) - follow the instructions in the preprocessing folder

Run the code

To run the benchmark script, use the following commands: bash conda activate gtamp export PYTHONPATH=$PYTHONPATH:`pwd` python scripts/01_benchmark.py -task_name shelf -task_id 1 -robot_name panda -planner multi_contact -res_file data/res.pkl python scripts/01_benchmark.py -task_name shelf -task_id 1 -robot_name ur5 -planner multi_contact -res_file data/res.pkl

To run for a single task and visualise the policy, use the file script/02_solve_task.py

To visualize the quantitative results, run: bash python visualisations/paper_visualization.py -res_file data/res.pkl

Install PDDLStream

In order to run PDDL method in benchmark, follow next steps: - Install PDDLStream repo - add import of MutableSet to pddlstream/examples/pybullet/utils/pybullet_tools/utils.py - if you use python >= 3.10, fix Sequence, Iterator and Sized imports from collection in pddlstream/language/generator.py, pddlstream/language/stream.py, pddlstream/algorithms/instantiation.py, pddlstream/algorithms/skeleton.py - Pass the correct -pddl_path argument to scripts

Owner

  • Name: agimus-project
  • Login: agimus-project
  • Kind: organization

GitHub Events

Total
  • Push event: 21
  • Create event: 1
Last Year
  • Push event: 21
  • Create event: 1

Dependencies

.github/workflows/static_code_analysis_and_tests.yml actions
  • actions/checkout v3 composite
  • actions/setup-python v4 composite
poetry.lock pypi
  • black 23.3.0 develop
  • click 8.1.3 develop
  • colorama 0.4.6 develop
  • importlib-metadata 6.7.0 develop
  • mypy-extensions 1.0.0 develop
  • packaging 23.1 develop
  • pathspec 0.11.1 develop
  • platformdirs 3.8.0 develop
  • ruff 0.0.260 develop
  • tomli 2.0.1 develop
  • typed-ast 1.5.4 develop
  • typing-extensions 4.6.3 develop
  • zipp 3.15.0 develop
pyproject.toml pypi
  • python >=3.7