This is an approach to implement RRT's from scratch, RRT's allow for multi-dimensional path-planning and object-collision-detection.
The code is written entirely without numpy
, it simply uses a library which depends on numpy to visualize the tree.
Currently the code supports two modes:
- RRT: includes a root-node the tree grows from there until it finds the target area and returns the found path
- bi-RRT: two root nodes are places and an overlap in the branch returns the found path
The algorithm is based and implemented from Rapidly Expanding Random Trees original paper
- RRT* implementation
- make a library