KUKA Robotics Challenge

This repo contains ROS packages to move a payload through a 2D maze using a KUKA industrial robotic arm as shown below.

This project aims to solve the KUKA Robotics challenge organized by Udacity, KUKA, and KIT. This repo builds upon ROS packages provided by KIT, available on Github, here and here. Really good documentation about these packages is available on KIT's KUKA Robotics Learning Lab website here

The available code provides an easy interface using ROS services to communicate with the KUKA Robotic Arm. The path planning and control algorithm I implemented, used the following services:

• check_path: verify if a linear path between two 2D poses is valid (collision-free and within workspace)
• move: command the robot by sending a 2D pose, the robot will move on a linear path to this pose

Setup

This project was tested with ROS Kinetic.

A catkin workspace can be setup as follows:

mkdir ~/catkin_ws
cd ~/catkin_ws
catkin_init_workspace
mkdir src
cd src
git clone https://github.com/ssharma1991/kuka-robotics-challenge

To run the project, use the created shell script as follows:

cd rll_planning_project/scripts
./start_project.sh

The script starts all required ROS launch files to start the simulation in Gazebo and visualization in Rviz. The payload dimensions, start pose and final pose can be tweaked within planning_iface.launch file.

Video of a complete simulation is available here: https://github.com/ssharma1991/kuka-robotics-challenge/tree/master/videos

Algorithm

A-star algorithm has been used for Path Planning. The state space is 3D and includes (x, y, theta) coordinates. A grid resolution of 2cm was used to discretize the maze and gave satisfactory results. The path planning code exists in path_planning.py file.