Light Curtain Constraint Graph

A planning and optimization library for light curtains using the constraint graph.
Efficient C++ implementation, with APIs for Python and Javascript!

Constraint Graph	Dynamic Programming

Installation

1. Clone repository

• Run the following commands:

  git clone --recursive /~https://github.com/CMU-Light-Curtains/ConstraintGraph.git
  cd ConstraintGraph

2. Make sure that MAX_RAYS and MAX_RANGES_PER_RAY are set to reasonable values

• Set them in include/planner/base.h.
• For correctness, they should always be greater than or equal to the true number of rays and ranges per ray.
• For efficiency, they should be as small as possible, ideally equal to the true rays and ranges per ray.

3. Build with python bindings

• Run the following commands:

  rm -rf build && mkdir build && cd build
  cmake -DCMAKE_BUILD_TYPE=Release \
        -DBIND_PY=ON \
        -DPYTHON_EXECUTABLE=<<PATH-TO-PYTHON-EXECUTABLE>> ..
  make -j
  cd ..

• To test the python build, run the following commands. The output should match apis/py/output.txt.

  cd apis/py
  python lc_planner/planner.py

4. Build with javascript bindings

• Download and install emscripten: [Click here for download and install instructions]
• After installation, relaunch your shell, and make sure that the $EMSDK environment variable has been set to point to the directory where emscripten has been installed. Also make sure that the following file exists: $EMSDK/upstream/emscripten/cmake/Modules/Platform/Emscripten.cmake
• Run the following commands:

  rm -rf build && mkdir build && cd build
  cmake -DCMAKE_BUILD_TYPE=Release \
        -DBIND_JS=ON ..
  make -j
  cd ..

• To test the javascript build, run the random curtain demo.
  • Launch the demo's HTTP server using the following command:

    python -m http.server -d apis/js/random-curtain-demo

  • Run the URL http://127.0.0.1:8000/ in your browser.