This repository builds feature-based EKF SLAM on Turtlebot3 from scratch. The core modules are:

• Odometry - odometry calculations are performed by Differential Drive Kinematics
• Measurement - both LiDAR and ArUco measurements are implemented
  • LiDAR measurement - unknown data association in lidar_ekf_slam package
  • ArUco measurement - known data association in aruco_ekf_slam package

Please navigate to the respective directories for a detailed description of each component.

• turtlelib - contains 2D Lie Group operations for Transforms, Vectors and Twists as well as differential drive robot kinematics for odometry updates
• nuturtle_description - houses the description of a differential drive robot with a caster wheel for support
• nuturtle_control - develops some nodes that are useful in both simulation and on the real robot
• nusim - includes functions of position tracking and obstacle setting
• nuslam - introduces LiDAR feature detection and EKF SLAM implementation

• Rviz Simulation
  • Youtube Link: https://youtu.be/ZHGrI5ZNQc0

• Real World Video
  • Youtube Link: https://youtu.be/6MdL9HDXfKs

• catkin_make - build the package
• roslaunch nuslam unknown_data_assoc.launch - launch LiDAR EKF SLAM algorithm with unknown data association

Dataset Link: https://pan.baidu.com/s/10crRfgGcZ-XgcBjefxAPDg Password: pqby

• The launch file (slam.launch) should be modified to the correct download path

• catkin_make - build the package
• roslaunch aruco_ekf_slam slam.launch - launch ArUco EKF SLAM algorithm with known data association
• rosbag play aruco_slam_data_qhd1.bag -r 5 - play the rosbag

Ubuntu20.04, OpenCV 3.1, Eigen, ROS

1. OpenCV library for the fundamental functions of Aruco codes: https://docs.opencv.org/3.3.0/d9/d6d/tutorial_table_of_content_aruco.html
2. Probalisitic Robotics -- by Thrun, S., Burgard, W., Fox, D. The MIT Press (2005)