Skip to content

Geometry generation/planning for robotically assembled spatial structures

License

Notifications You must be signed in to change notification settings

yijiangh/coop_assembly

 
 

Repository files navigation

coop_assembly

compas version compas_fab version
License MIT https://travis-ci.com/yijiangh/coop_assembly.svg?branch=dev Documentation Status https://coveralls.io/repos/github/yijiangh/coop_assembly/badge.svg?branch=dev

coop_assembly: Geometry generation of robotically assembled spatial structures.

News:

9-21-2019: This package contains materials for the FABRICATION-INFORMED DESIGN OF ROBOTICALLY ASSEMBLED STRUCTURES for the Design Modeling Symposium Berlin 2019.

Getting Started

See Getting Started for installation instructions.

Planning examples

Regression search examples

First, make sure that you first follow the Installating sequence and motion planning stack section of the Getting Started document.

Then, Run the regression search algorithm for find a construction plan for spatial bars.

python -m coop_assembly.planning.run -v -w

Issue python -m coop_assembly.planning.run -h to print out the helps.

And you should be able to see the following (press Enter when asked in the commandline) animation (click image for the youtube video):

bar assembly simulated demo

PDDLStream examples (WIP)

First, follow the Installating PDDLStream (WIP) section of the Getting Started document to set up PDDLStream and pyplanners.

2D version

This is a 2D additive construction domain, developed for testing our algorithms.

python -m examples.assembly2D.run -v -w -p 2D_tower_skeleton.json -a incremental_sa

The incremental_sa flag is the incremental algorithm in PDDLStream, using the pyplanners for supporting semantic attachments.

Issue python -m examples.assembly2D.run -h to print out the helps.

If run correctly, you should see an animation (click image for the youtube video) like the following (subject to changes) after a plan is found:

bar assembly simulated demo

3D version

WIP.

Troubleshooting

Sometimes things don't go as expected. Checkout the troubleshooting documentation for answers to the most common issues you might bump into.

Credits

This package was initiated by Stefana Parascho <[email protected]> @stefanaparascho at the CREATE lab, Princeton University. The sequence and motion planning scripts are developed by Yijiang Huang <[email protected]> @yijiangh with collaborators.

Releases

No releases published

Packages

No packages published

Languages

  • Python 77.3%
  • Jupyter Notebook 18.8%
  • Julia 3.9%