Motion Planning Algorithms for Non Holonomic Robots using ROS and Turtlebot

Overview

This repository implements different motion planning algorithms for non holonomic robots using ROS. All the algorithms are tested using turtlebot in Gazebo.

Non Holonomic Robot

In the previous project, holonomic robot behaviour was implemented wherein the robot could move sideways. Vehicles (cars, mobile robots etc) are modelled as non-holonomic robots where a neighbour coordinate along a trajectory given the position and orientation is expressed as -

$$ \begin{align} \begin{split} x_{pos} &= x_{pos} + \frac{radius \ast (u_L + u_R) \ast cos(\theta) \ast dt}{2} \\ y_{pos} &= y_{pos} + \frac{radius \ast (u_L + u_R) \ast sin(\theta) \ast dt}{2} \\ \theta_{initial} &= \theta_{initial} + \frac{radius \ast (u_R-u_L) \ast dt}{L} \end{split} \end{align} $$

Where $u_L$ and $u_R$ are the velocity of left and right wheel, $radius$ is the radius of the wheel, $L$ is the distance between the centre of two wheels.

Project Structure

    .
    ├── Search Based            
    │   ├── A star       
    │   └── Dijkstra       
    |    
    ├── Curve Based            
    │   ├── Dubins Curve       
    │   └── Reed Shepp Curve       
    |
    └── Sampling Based
        ├── RRT       
        └── RRT*            

A* Algorithm

RRT

Dubins Curve

Contact

If you have any questions, please let me know:

• Shaswat Garg {[email protected]}