Robotics Project

NUS, EE4308, AY19/20, Semester 2

Authors: Marc Rauch*, Jan Hünermann*

This is the code for the robotics project of EE4308 at NUS. The goal of the project was having the robot navigate an unknown environment on its own. For that, we implemented

• an odometry motion model
• binary log-odds model for lidar based mapping
• Theta-Star global path planner
• Spline based local planner with pure pursuit
• PID controller

Furthermore, in the second part of the project we implemented an extended Kalman filter for controlling a drone that is able to fly between the target and robot position.

*: Equal contribution

Install

Make sure you have the following requirements met:

• ROS Melodic
• Eigen 3 for math routines
• OpenCV 3 or higher
• Gazebo 9 with TurtleBot
• hector_quadrotor

To install, clone this repository inside a folder called catkin_ws/src. Make sure you have all your other workspaces sourced, including the workspace containing the hector_quadrotor packages. Then build the workspace by running the following command from the directory catkin_ws.

catkin_make install -DCMAKE_BUILD_TYPE=Release

Launch

To launch the robot, run the following command:

roslaunch bringup robot_bringup.launch

or

roslaunch bringup proj1.sh

to start the mission.

TODO

Part 1

• Mapping and motion model
• Global planning
• Trajectory generation and following

Part 2

• Sensor fusion
• Extend mission planner
• Motion control (easy)
• Modify launch structure
• Fine-tune ground robot

Topics

• /robot_pose: geometry_msgs/Pose2D
• /map: nav_msgs/OccupancyGrid
• /navigation/goal: geometry_msgs/Pose2D
• /navigation/path: nav_msgs/Path, global coordinates
• etc