Skip to content

ROS control

This page describes the ROS control for the MiR100 robot only . To control the mobile base as part of the MiRco mobile manipulator see MiRco ROS control.

To establish a ROS connection with MiR100 see Connect to MiR100.

Warning

MiR100 internal clock is prone to desynchronization which can mess with the robot control.
Before controlling the robot, you have to connect to the MiR web interface and go to System -> Settings -> Date & Time and synchronize the internal clock. Click Load from device and Save changes.

Note

For an advanced setup for clock synchronization see this solution.

Packages overview

  • mir_control: main control package
  • mir_rest_api: MiR REST API
  • mir_robot: MiR ROS driver developed by the DFKI team
  • mir_joy_teleop: joystick control

Usage

Start mir_control

# MiR100 connection over hotspot
roslaunch mir_control mir_control.launch

# MiR100 connection over outside network
roslaunch mir_control mir_control.launch mir_ip:=<MiR100_IP>

RViz

After you started mir_control, you can send the robot to a goal using the 2D Nav Goal tool in RViz.

Make sure to set these RViz settings:

Panels -> Tool Properties -> 2D Nav Goal -> Topic: /mir100/move_base_simple/goal
Panels -> Tool Properties -> 2D Pose Estimate -> Topic: /mir100/initialpose
Global Options -> Fixed Frame: map

Send the robot to a goal:

  • use 2D Pose Estimate to mark the approximate MiR100 start location
  • use 2D Nav Goal to send the robot to a goal

mir_control package

The main control package. This is a wrapper around the mir_driver package. For details about mir_driver see the GitHub page.

mir_control starts another roscore on the host computer, connecting to the MiR100 internal roscore over ROS Bridge. See the MiRco ROS schematic for a visual aid of the connection.

mir_control adds a control node, a python robot class, a goal teacher, ROS services and REST API integration. This effectively enables the use of all web interface functions.

Launch files

mir_control.launch

  • launches mir_driver
  • launches mir_control_node

Nodes

mir_control_node.py

The main control node for MiR100. Initializes a MiR100 python object. See mir100_class.py script for details on the object.

Config files

target_goals.yaml

When you run the control node, the positions in the current MiR map are saved in this file.
If taget_goals.yaml already contains a position with the same name (this file is not reset on start), the old position is overwritten.

# saved positions structure
position_name:
  orientation:
    w: 0.0112920402716304
    x: 0.0
    y: 0.0
    z: -0.9999362428807668
  position:
    x: 21.218
    y: 13.802
    z: 0      # z coordinate is always saved as 0

markers.yaml

When you run the control node, the markers in the current MiR map are saved in this file.
This doesn't include marker entry positions. Currently only VL type markers are supported.
If markers.yaml already contains a marker with the same name (this file is not reset on start), the old marker is overwritten.

# saved markers structure
marker_name:
  guid: 017f3480-fb3d-11ee-947c-94c691a73828
  offsets_guid: 018037c1-fb3d-11ee-947c-94c691a73828
  orientation_offset: 2.0
  type_id: 11
  x_offset: -0.6
  y_offset: 0.0

Services

Services are established with the mir_control_node or whenever you create a MiR100 object defined in mir100_class.py.

save_mobile_goal (TODO: check if works correctly)

Saves the current MiR100 pose under selected name. The pose is saved both in the web interface and in /config/target_goals.yaml.

If the current MiR100 map does not include a position or marker with the same name, a position with the selected name is created and saved to the current MiR map. If a position or marker with the same name already exists, they are overwritten with the current MiR100 pose.

send_to_goal

Send robot to a goal saved in /config/target_goals.yaml.

get_markers

Get list of markers saved in /config/markers.yaml.

dock_to_vl_marker

Start docking to the selected marker saved in config/markers.yaml.

change_marker_offsets

Change the offsets of a marker saved in /config/markers.yaml.
Currently only VL type markers are supported.

mir_joy_teleop package

A package for joystick control.

As with connecting to the robot, you can send the joystick commands directly to the MiR100 internal roscore or use it with the mir_control package.

Note

When using mir_control we send geometry_msgs/Twist type messages instead of geometry_msgs/TwistStamped to the /cmd_vel topic. That is because the mir_driver package expects messages of type geometry_msgs/Twist on the /cmd_vel topic and converts them to geometry_msgs/TwistStamped messages before sending the commands to the robot.

Internal

Follow the Establish ROS connection - internal roscore instructions.
Then run:

# launch a joy_node and a teleop node
roslaunch mir_joy_teleop joy_teleop.launch

# default input device is js1
# you can specify input device - e.g. js2
$ roslaunch mir_joy_teleop joy_teleop.launch device:=js2

mir_control package

Follow the Establish ROS connection - mir_control package instructions.
Then run: 

# launch a joy_node and a teleop node
roslaunch mir_joy_teleop joy_teleop.launch roscore:=external

# default input device is js1
# specify input device - e.g. js2
$ roslaunch mir_joy_teleop joy_teleop.launch device:=js2 roscore:=external