Tracebot Mockup

This repository contains ROS packages to run a simulated mockup of the Tracebot setup in Rviz and Gazebo. The objective is to analyse different placements of the robots and equipment in regard to reachability and manipulability.

Table of Contents

• Model Parameters
• Usage
  • Launch Rviz
  • Launch Gazebo
• Setup
  • Installing locally
  • Using docker images

Model Parameters

The URDF model contained in tracebot_mockup_description is parametrized using xacro. The table below lists the available parameters and their meaning.

Parameter name	Description
arm_mount_offset_x	Offset in X direction of the robot arm with respect to the center of the stand.
arm_mount_offset_y	Offset in Y direction of the robot arm with respect to the center of the stand.
arm_mount_offset_z	Offset in Z direction of the robot arm with respect to the top of the stand.
arm_mount_offset_theta	Rotational offset around Z of the robot arm with respect to the center of the stand.
mount_base_height	Height of the robot stand.
mount_base_length	Length of the robot stand.
mount_base_width	Width of the robot stand.
arm_base_tilt	Angular upward tilt of the robot bases.
left_arm_model	Model used for left Robot Arm. Must use on of: (ur3, ur3e, ur5, ur5e, ur10, ur10e, ur16e)
right_arm_model	Model used for right Robot Arm. Must use on of: (ur3, ur3e, ur5, ur5e, ur10, ur10e, ur16e)

All parameters use SI units.

Usage

Launch Rviz

This repository offers visualisation of the mockup model within rviz using the view_tracebot_mockup.launch file, inside tracebot_mockup_description.

The rviz simulation give a basic visualisation of the robot with any of the robot arms (the default being ur5e). The launchfile exposes the parameters listed in Model Parameters as arguments, providing reasonable defaults.

Firsly ensuring that the workspace is sourced:

source ~/path/to/tracebot_mockup_ws/devel/setup.bash

• The mockup can be run with default parameters:

  roslaunch tracebot_mockup_description view_tracebot_mockup.launch

• Or using any of the available parameters, such as exchanging the left arm for the ur10e model:

  roslaunch tracebot_mockup_description view_tracebot_mockup.launch left_arm:="ur10e"

A camera view can be added to rviz:

roslaunch tracebot_mockup_description camera_display.launch

The intrinsic camera parameters are defined using the standard ROS format, and stored in head_camera.yaml. The pose of the camera in the world is defined within the launchfile camera_display.launch

Launch Gazebo

Parameter name	Description
controller_config_file	Config file used for defining the ROS-Control controllers.
controllers	Controllers that are activated by default.
stopped_controllers	Controllers that are initally loaded, but not started.
tf_prefix	tf_prefix used for the robot.
tf_pub_rate	Rate at which robot_state_publisher should publish transforms.
paused	Starts Gazebo in paused mode
gui	Starts Gazebo gui

As shown above, the gazebo launch file has some additional parameters that can be set. However any of the previously stated parameters can also be used.

• This can be launched with the default using:

  roslaunch tracebot_mockup_gazebo view_tracebot_gazebo.launch

• Or using any of the available parameters, such as starting the simulation paused:

  roslaunch tracebot_mockup_gazebo view_tracebot_gazebo.launch paused:=true

The arms can then be communicated with by publishing to the /pos_joint_traj_controller/command topic which uses JointTrajectory messages. An example of this can be found in the scripts/gazebo_model_test.py file. To run this, and test that the gazebo simulation is working, run:

rosrun tracebot_mockup_gazebo gazebo_model_test.py

You should see that all the motors trigger and the arms curl up.

Setup

There's two different possibilities to use the packages in this repository: installing locally or running the pre-built docker images.

Installing locally

These packages are known to support ROS Melodic and Noetic, although other distros may be supported as well.

The packages can be built/installed following standard ROS-based procedures. The examples listed below use catkin_tools as build tool, although catkin_make and catkin_make_isolated should be also supported:

• Create a workspace:

  mkdir -p ~/path/to/tracebot_mockup_ws/src
  cd ~/path/to/tracebot_mockup_ws
  catkin config --extend /opt/ros/"$ROS_DISTRO"

• Pull the packages into the repository:

  cd ~/path/to/tracebot_mockup_ws/src
  git clone https://gitlab.com/tracebot/tracebot_mockup.git

• Certain dependencies are not released as binary packages to either Melodic or Noetic, pull those into the workspace as well:

  cd ~/path/to/tracebot_mockup_ws/src
  vcs import < tracebot_mockup/upstream.repos # Install python(3)-vcstool if not available

• Install the rest of dependencies from binary repositories

  cd ~/path/to/tracebot_mockup_ws
  rosdep install -iy --from-paths src --rosdistro "$ROS_DISTRO"

• Build:

  cd ~/path/to/tracebot_mockup_ws
  catkin build

Using docker images

The miguelprada/tracebot_mockup:noetic docker image with the mockup packages pre-installed is periodically built and made available through DockerHub.

The recommended way to run this image is using the osrf/rocker tool. This is a small wrapper tool around the docker CLI which facilitates tasks such as enabling X11 forwarding (required to view the simulated scene in rviz).

The mockup with default parameters can be run using rocker with:

rocker --x11 -- miguelprada/tracebot_mockup:noetic roslaunch tracebot_mockup_description view_tracebot_mockup.launch

Note that if your system runs NVidia graphics, you may need to use:

rocker --x11 --nvidia -- miguelprada/tracebot_mockup:noetic roslaunch tracebot_mockup_description view_tracebot_mockup.launch

Alternatively, instructions to enable X11 forwarding in docker in different ways can be found in this page.