/rrt_toolbox

RRT, RRT*, RRT*FN algorithms for MATLAB

Primary LanguageMATLABBSD 3-Clause "New" or "Revised" LicenseBSD-3-Clause

DOI

ko-fi

MATLAB implementation of RRT, RRT* and RRT*FN algorithms.

What is RRT, RRT* and RRT*FN

  • RRT (Rapidly-Exploring Random Tree) is a sampling-based algorithm for solving path planning problem. RRT provides feasable solution if time of RRT tends to infinity.

  • RRT* is a sampling-based algorithm for solving motion planning problem, which is an probabilistically optimal variant of RRT. RRT* converges to the optimal solution asymptotically.

  • RRT*FN is a sampling-based algorithm based on RRT*. RRT*FN inherents asymptotical convergence to the optimal solution, however RRT*FN implements it using less memory.

How to use

The original package contains 3 files containing algorithm

  • rrt.m
  • rrt_star.m
  • rrt_star_fn.m

and 2 files containing classes that enable algorithm to solve problems for simple 2D mobile robot model and nDOF Redundant Manipulator

  • FNSimple2D.m
  • FNRedundantManipulator.m

One can add other additional models implementing all methods mentioned in rrt.m, rrt_star.m, rrt_star_fn.m

Authors

  • Olzhas Adiyatov oadiyatov[at]nu.edu.kz
  • Atakan Varol

(c) 2013