Motion Planning

This repository is used for implementing and testing motion planning algorithms. The package includes several algorithms such as PRM (Probabilistic Roadmap) and Bidirectional RRT (Rapidly-exploring Random Tree). Following OOP principles, the algorithms are implemented in the way that can be reusable and further developed in the future.

Algorithms

• PRM (Probabilistic Roadmap): A sampling-based algorithm that constructs a roadmap of the free space. It uses Dijkstra's algorithm to find the shortest path from the generated roadmap.
• Bidirectional RRT (Rapidly-exploring Random Tree): An algorithm that grows two trees from the start and goal points and attempts to connect them.

Side Notes

• Testing Only: This package is designed for testing algorithms only. The environment and map are included inside the package.
• Not Suitable for Real Projects: The algorithms in this package are not yet suitable for integration with real projects.
• Future Work: To make the package integrable and reusable as a separate library, the algorithms can be turned into ROS action servers. ROS topics or services can be used to enable communication within the system.

Usage

To reconfigure the map as well as the algorithms, all you need to do is adjusting the value of parameters in YAML files inside cfg directory. For example, to select the algorithm, please navigate to cfg/map.yaml and change the value of algorithm, currently supporting 'prm' and 'rrt'.

To run the motion planning algorithms, use the provided launch files launch.launch and configure the parameters as needed. The launch files launch.launch load the necessary parameters and start the main node, to launch, please run this command line (after starting roscore):

roslaunch motion_planning launch.launch

Configuration

The package includes several configuration files in the cfg directory:

• map.yaml: Configuration for the map.
• post_processing.yaml: Configuration for post-processing steps.
• prm.yaml: Configuration for the PRM algorithm.
• rrt.yaml: Configuration for the RRT algorithm.

In the src directory, the following files are included:

• prm.py: Implements the PRM algorithm, generating samples and graphs that can be used with other graph-based pathfinding algorithms such as BFS and Dijkstra's.
• rrt.py: Implements the Bidirectional RRT algorithm on the given map.
• post_processing.py: Applies post-processing techniques to the generated path, currently supporting path-shortcutting.
• main.py: Contains the main pipeline to process PRM or RRT with post-processing, reading configurations from the cfg directory.

The utils directory contains essential components for implementing and testing the algorithms, including environment setup, map handling, and search algorithms.

Results

Here are some results of the motion planning algorithms:

PRM Algorithm	PRM Algorithm with Post-Processing

Bidirectional RRT Algorithm	Bidirectional RRT Algorithm with Post-Processing

Dependencies

• numpy
• pylab
• rospy

Make sure to install the required dependencies before running the package.