planning-algorithms

Experimental implementations of Planning Algorithms by Steven LaValle.

Installation

git clone [email protected]:mech0ctopus/planning-algorithms.git

## C++
cd planning-algorithms/modules/cpp
mkdir build && cd build
cmake ..
make

## Python
cd planning-algorithms/modules/python3
# Install in editable mode (don't forget the period at the end)
pip3 install -e .

Tests

## Run all tests
cd planning-algorithms
./run_tests.sh

## C++
cd planning-algorithms/modules/cpp/build
./TestPrimitives

## Python
cd planning-algorithms/modules/python3
nosetests3 .

Examples

cd planning-algorithms/modules/python3/examples
python grid2d.py
python grid3d.py
python five_state.py
python optimal_grid2d.py
python optimal_grid3d.py
python optimal_five_state.py

TODO

• Update unit tests for ForwardLabelCorrecting

TODO Later

• Optimal Planning with Forward Value Iteration

Maybe Later

• Implement:
  • A-Star
  • IterativeDeepening
• Define a better SearchProblem interface
• Mark Initial/Goal States as VISITED in bidirectional search
• Create/Implement more examples
• Determine a path forward for C++ work
• Identify a better design pattern to use for constructing Forward/Backward versions of the same search algo. (Factory?)
  • Should Forward, Backward, and Bidirectional search be consolidated into a Unified interface?
• Add plot results to problem class
• Raise exception if search fails (instead of Return code)
• Add logic to ForwardSearch for checking if a state is alive or dead per p.33