vshkar / slam_project Goto Github PK

"ME495 Sensing, Navigation, and Machine Learning" Author : Vishwajeet Karmarkar

This repository houses code for a SLAM class at Northwestern Univerisity, January 2020.

├── Homework2_Answers.md
├── nuslam
├── nuturtle_description
├── nuturtle_gazebo
├── nuturtle_robot
├── README.md
├── rigid2d
├── SLAM_RESULTS.md
└── tsim

The structure is shown above.

1. rigid2d: A c++ library for modelling a diff drive robot and kinematics. Also houses helper functions.
   
2. nuturtle_description: Urdf files, launch files for starting Rviz visualisation
   
3. nuturtle_robot: Robot control code
4. nuturtle_gazebo: Xacro files along with sensor plugins for Gazebo simulation
   
5. tsim : Waypoint follower for turtlebot
   
6. nuslam: Slam code

To run:

mkdir src 
cd src 

git clone https://github.com/vishwajeet-NU/slam_project.git

catkin_make

To launch slam

You can launch it in debug mode and normal mode
In debug mode gazebo model state is used to populate sensor messages
so that slam can be tested without worrying about data association

to launch only landmark detector::

roslaunch nuslam landmarks.launch robot:=-1 

It looks as follows

Robot argument is dependent on choice of platform
this launch file also starts a keyboard teleop system which can be used to
navigate the bot

to launch slam

roslaunch nuslam slam.launch robot:=-1 debug:=True 

debug argument can be changed as needed