2D Rapidly-exploring Random Tree (RRT) for Robot Path Planning

Author: Muye Jia

This RRT algorithm allows for 2D path planning for robots. The tree can find a free path around obstacles with any shapes and sizes, and the path along with the Tree generated is visualized using Matplotlib.

Input a map image file with 1 representing obstacles and 0 representing free space. The output would be a path stored as csv file named path.txt and a mp4 file ani.mp4 with planning animation.

Example

The algorithm is going to plan a path around the Northwestern-N-shaped obstacle. Starting at (40, 40), ending at (60, 60). The resulted path (plot in red):

The RRT expansion animation:

n_map.mp4

Quick Start Guide

1. To run the algorithm with customized obstacle map (or the N_map.png in the repo), run the following command: python3 obstacle_avoidance.py <map_name> <tree_step_size> <planning_domain> <max_node_number> <start_x_position> <start_y_position> <goal_x_position> <goal_y_position>. Example run: python3 obstacle_avoidance.py N_map.png 1 100 500 5 5 40 40.

2. Note: The input map file needs to be in 1 and 0, where 1 represents the obstacle in map and 0 represents free space.

Required Packages

Need FFMpeg for Matplotlib animation.