This is the or_tf plugin for OpenRAVE. This plugin is listening to tf frames and update OpenRAVE Kinbodies in realtime. It can be used with any perception system that publishes tf to the ROS's tf topic.

This is a catkin package, which means that you need to configure your catkin workspace. If you haven't configured your workspace yet, then please refer to this tutorial.

Prerequisites

• You need to install the openrave_catkin package by just cloning the repository in your ~/catkin_ws/src.
• Optionally you might also need mocap_optitrack or anything else applicable to your own configuration. mocap_optitrack is the one which publishes the tfs to ROS in our configuration for example, but has nothing to do with or_tf plugin.

Installation

Installation is as simple as cloning this repository in your ~/catkin_ws/src and then running catkin_make within ~/catkin_ws.

cd ~/catkin_ws/src
git clone [email protected]:papallas/or_tf.git
cd ~/catkin_ws
catkin_make

You should see a success message at the end of this.

Let OpenRAVE know about the new plugin

You need to include the following line in your ~/.bashrc file after you run catkin_make:

export OPENRAVE_PLUGINS=$OPENRAVE_PLUGINS:~/catkin_ws/devel/share/openrave-0.9/plugins
export PYTHONPATH=$PYTHONPATH:~/catkin_ws/src/or_tf/pythonsrc/or_tf

Where openrave-0.9 is the version of your OpenRAVE.

The first line is appending the path where or_tf is being built so OpenRAVE is aware of its path, where the second is letting Python know about the wrapper class so you can import it into any script and use the wrapper class.

Make sure that you either restart your terminal window or you run source ~/.bashrc.

Moreover, verify the success of the installation and configuration by running openrave --listplugins and you should see the following output:

sensor: ...
.
.
.
or_tf - /.../catkin_ws/devel/share/openrave-0.9/plugins/or_tf_plugin.so
.
.
.

Setting up your script and registering OpenRAVE kinbodies to tf frames

Your code requires the following import:

from or_tf import OpenRAVE_TF

And also the following code to initialise the or_tf wrapper class:

# Create or_tf instance with frame id.
or_tf_plugin = OpenRAVE_TF(env, "world")

You now need to get our OpenRAVE kinbody:

# Get Kinbody from the environment.
object_1 = env.GetKinBody("object_1")

And you need to register the kinbody to the tf plugin:

or_tf_plugin.RegisterBody(object_1, "goal/base_link")

How you will find the goal/base_link parameter?

Run rosrun tf tf_monitor and you should get a similar output to this one:

RESULTS: for all Frames

Frames:
Frame: base published by unknown_publisher Average Delay: 7.94464 Max Delay: 7.96316
Frame: base_link published by unknown_publisher Average Delay: 7.94463 Max Delay: 7.96316
Frame: ee_link published by unknown_publisher Average Delay: 7.94463 Max Delay: 7.96316
Frame: forearm_link published by unknown_publisher Average Delay: 8.44579 Max Delay: 8.45808
Frame: goal/base_link published by unknown_publisher Average Delay: 2.97504 Max Delay: 2.9932
Frame: robot/base_link published by unknown_publisher Average Delay: 2.97725 Max Delay: 3.00143
Frame: shoulder_link published by unknown_publisher Average Delay: 8.4458 Max Delay: 8.45808
Frame: tool0 published by unknown_publisher Average Delay: 7.94463 Max Delay: 7.96316
Frame: tool0_controller published by unknown_publisher Average Delay: 8.44492 Max Delay: 8.45965
Frame: upper_arm_link published by unknown_publisher Average Delay: 8.4458 Max Delay: 8.45808
Frame: wrist_1_link published by unknown_publisher Average Delay: 8.4458 Max Delay: 8.45808
Frame: wrist_2_link published by unknown_publisher Average Delay: 8.4458 Max Delay: 8.45808
Frame: wrist_3_link published by unknown_publisher Average Delay: 8.4458 Max Delay: 8.45808

From there you can see the goal/base_link which is the one you need. The name goal is coming from your configuration and is subject to your setup. For example if you are using OptiTrack and you are using the ROS Package mocap_optitrack then the goal will be whatever you have named your frame in the config/mocap.yaml file:

#
# Definition of all trackable objects
# Identifier corresponds to Trackable ID set in Tracking Tools
#
rigid_bodies:
    '1':
        pose: goal/pose
        pose2d: goal/ground_pose
        child_frame_id: goal/base_link
        parent_frame_id: world
optitrack_config:
        multicast_address: 224.0.0.1

where 1 is the user_id in OptiTrack.

A testplugin.py file is provided as an example.

Contributors

or_tf plugin was developed by the Robot Manipulation Lab in the School of Computing at the University of Leeds. This software is initially developed by Dr. Mehmet Dogar and is currently maintained by Rafael Papallas.