This is a dynamic smoothed-RRT planner for Lynx robot(6-DoF manipulator).

• Main functions:

  • Simulation function: runsim.m
  • Generative function for static planning(main function): SRRT.m
  • Generative function for dynamic planning: regrow.m

• Other functions:

  • Sample function: sample.m
  • Pick random node in space: RandomNode.m
  • Neighbor finding function: neighbor.m
  • Node extending function: extend.m
  • Collision detection function: DetCol.m
  • Path optimization function: path_opt.m

• Utils: Those are helper functions either from p-code or m-code that I implemented in the former labs

To evaluate the smoothness of the planner, we simulated in different static maps:

• static simulation results:

  • comparison between random sample and epsilon-greedy sample:

  • comparison between raw and pruned path:

    

To evaluate the dynamic performance of the planner, due to the fact that visualizing 3d moving obstacle in matlab is quite tricky, we implement it onto a planner dot robot navigating in changing environment and get the following result:

• scenario 1: randomly moving door:

• scenario 2: tricky maze: