UTD University Rover Challenge

This is where all of our code and most of the project management will be base at.

Overview

Discussions: Where all research topics will go and where you will log your ideas, progress, and questions at. I will also post research materials there too.

Issues: Once research is done we will create an issue for a topic and this is where you will contribute your code too and discuss issues that are arriving in the implementation.

Instructions

• Start up the simulation and press play.

• In another terminal, cd to rtab_package (such a dumb name we can definitely change this) which is in the ws_urc src folder, colcon build, source the install.

• To run, "ros2 launch rtab_package rtab.launch.py".

  • This node takes care of all transforms needed for the Nav2 stack.
  • rgbd odometry creates an odom -> base_link odometry message
  • ekf robot_localization fuses the IMU odometry with the visual odometry and actually publishes the odom -> base_link transformation
  • rgbd-slam publishes the map -> odom frame
  • Currently uses the Frame-to-Frame model, there are more computationally expensive but robust options out there

• Then, cd to navigation, which is also in the ws_urc src folder. Colcon build, source the install.

  • Run with "ros2 launch navigation nav.launch.py"
  • This initializes the nav2 stack, nothing outputs after initialization unless a goal is given

• Start rviz2

  • Set the global frame to "map"
  • include "Robot Description" topic to see the robot (the wheels dont move with it even though the transforms are being published, don't know what's going on there)
  • Include "odometry" topic, set ros2 topic to "odometry/filtered" to see the fused odometry
  • Include "/map" topic (this doesn't show up until nav2 is activated)
  • Include "goal pose" to see the given goal
  • To send actions to the robot, in rviz, select 2d goal pose. Place this somewhere on the map, and the rover will attempt to navigate towards it.