A communication framework for robotics, automation, and the Internet of Things
See the Getting Started Guide!
See https://github.com/robotraconteur/robotraconteur/wiki/Documentation for documentation.
See the examples directory for examples in various programming languages.
- I found a bug! Please leave an issue on the GitHub Issues page
- I have a specific question about how to use Robot Raconteur: Please leave a question on GitHub Discussions Q&A
- I have a general question or comment: Please leave a message on GitHub Discussions.
The Quick Start example demonstrates the basic functionality of Robot Raconteur be creating a service, and then calling the service using a client. This example uses the "Reynard the Robot" Python package, which provides a simple cartoon robot.
Before running the example, make sure to install the required packages:
python -m pip install robotraconteur reynard-the-robot
On Linux, use python3
instead of python
to run the Python 3 interpreter. Use python3
in the rest
of the examples as well.
reynard_quickstart_service.py
import RobotRaconteur as RR
RRN = RR.RobotRaconteurNode.s
import threading
import reynard_the_robot
# Define the service definition for the quickstart service
reynard_quickstart_interface = """
service experimental.reynard_quickstart
object ReynardQuickstart
function void say(string text)
function void teleport(double x, double y)
end
"""
# Implement the quickstart service
class ReynardQuickstartImpl(object):
def __init__(self):
self.reynard = reynard_the_robot.Reynard()
self.reynard.start()
self._lock = threading.Lock()
def say(self, text):
with self._lock:
self.reynard.say(text)
def teleport(self, x, y):
with self._lock:
self.reynard.teleport(x, y)
with RR.ServerNodeSetup("experimental.minimal_create2", 53222):
# Register the service type
RRN.RegisterServiceType(reynard_quickstart_interface)
reynard_inst = ReynardQuickstartImpl()
# Register the service
RRN.RegisterService("reynard", "experimental.reynard_quickstart.ReynardQuickstart", reynard_inst)
# Wait for program exit to quit
input("Press enter to quit")
To run the service, execute the following command:
python reynard_quickstart_service.py
And open a browser to http://localhost:29201 to view the Reynard user interface.
This service can now be called by a connecting client. Because of the magic of Robot Raconteur, it is only necessary to connect to the service to utilize its members. In Python and MATLAB there is no boilerplate code, and in the other languages the boilerplate code is generated automatically.
reynard_quickstart_client.py
from RobotRaconteur.Client import *
# RRN is imported from RobotRaconteur.Client
# Connect to the service.
obj = RRN.ConnectService('rr+tcp://localhost:53222/?service=reynard')
# Call the say function
obj.say("Hello from Reynard!")
# Call the teleport function
obj.teleport(100, 200)
To run the client, execute the following command:
python reynard_quickstart_client.py
The MATLAB Add-On for Robot Raconteur can be installed using the Add-On Explorer in MATLAB and searching for "Robot Raconteur".
In MATLAB, the example client is even simpler.
reynard_quickstart_client.m
% Connect to the service
o = RobotRaconteur.ConnectService('rr+tcp://localhost:53222/?service=reynard');
% Call the say function
o.say("Hello from MATLAB!");
% Call the teleport function
o.teleport(-150,200);
The quickstart file can be found in the examples/quickstart directory.
Robot Raconteur has a large ecosystem with a number of related projects. Start with the Robot Raconteur Getting Started Guide.
The Python examples support using a Gazebo simulated iRobot Create robot. See the training simulator for installation instructions:
https://github.com/robotraconteur-contrib/robotraconteur_training_sim
The training simulator also contains a simulation for two Universal Robots UR5e robots, with grippers and cameras. Example scripts to control the robots are included. See the training simulator readme for instructions.
Numerous examples can be found in the examples/
Robot Raconteur provides a large number of standardized types to use with robots and other devices. See the standard robdef repository:
https://github.com/robotraconteur/robotraconteur_standard_robdef
There are numerous other projects, drivers, and resources in the ecosystem. See the directory for a full list of available resources:
https://github.com/robotraconteur/robotraconteur-directory
Also checkout the PyRI Open source teach pendant: https://github.com/pyri-project/pyri-core
See docs/common/installation.md for installation instructions.
The following platforms are supported:
- Windows (x86, amd64, arm64): C++, Python, C#, Java, MATLAB, LabView
- Linux (x86, x64, armhf, arm64): C++, Python, C#, Java, MATLAB, LabView
- MacOS (x64, arm64): C++, Python, C#, Java, MATLAB
- Android (x86, x64, armhf, arm64): C++, Java
- iOS (arm-v7, arm64): C++
- Browser (Chrome, Firefox, Edge, Safari): C++, Python, JavaScript
- FreeBSD (x64)
See docs/common/building.md for build instructions.
Robot Raconteur is available in ROS Noetic and ROS Humble using the robotraconteur
package. These packages are built
using the ros-noetic
and ros2-humble
branches. The ros2-humble
branch should work with other versions of
ROS 2, but swig version 4.0.2 or greater must be installed first.
A Robot Raconteur to ROS 2 bridge is available, allowing access to ROS 2 topics and services from Robot Raconteur:
https://github.com/robotraconteur-contrib/robotraconteur_ros2_bridge
The Robot Raconteur project has defined a number of standard service definitions that contain numerous structure, pod, namedarray, and object types. These types cover a range of common data types, and provide standardized interfaces to devices. These types should be used whenever possible so that services will be interoperable. The standard service types are available in the https://github.com/robotraconteur/robotraconteur_standard_robdef GitHub repository.
The Robot Raconteur Companion libraries are provided to assist in using the standard service types, along with other generic utility functions. Currently, the companion libraries contain the standard service types, info file loaders, and general utility functions. The following libraries are available:
- Python: https://github.com/robotraconteur/robotraconteur_companion_python
- C++: https://github.com/robotraconteur/robotraconteur_companion
- C#: https://github.com/robotraconteur/RobotRaconteurNET.Companion
The Python companion library can also be installed using pip install RobotRaconteurCompanion
See https://github.com/robotraconteur-contrib/robotraconteur_camera_driver/blob/master/robotraconteur_camera_driver.py for an example utilizing standard types and the companion library.
The Robot Raconteur project maintains a list of available drivers. The directory can be found here:
https://github.com/robotraconteur/robotraconteur-directory
Robot Raconteur Core is a ROS Quality Level 2 package. See the Quality Declaration for more details.
Contributors must sign a Contributor License Agreement (CLA). Please see https://www.wasontech.com/contributors to complete and return a signed agreement. Wason Technology, LLC uses the Harmony CLA (https://www.harmonyagreements.org/).
The Robot Raconteur core library is Apache 2.0 licensed.
"Robot Raconteur" and the Robot Raconteur logo are registered trademarks of Wason Technology, LLC. All rights reserved.
Robot Raconteur is covered United States Patent No. 10536560
Robot Raconteur is developed by John Wason, PhD, Wason Technology, LLC
This work was supported in part by Subaward No. ARM-TEC-18-01-F-19 and ARM-TEC-19-01-F-24 from the Advanced Robotics for Manufacturing ("ARM") Institute under Agreement Number W911NF-17-3-0004 sponsored by the Office of the Secretary of Defense. ARM Project Management was provided by Christopher Adams. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of either ARM or the Office of the Secretary of Defense of the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes, notwithstanding any copyright notation herein.
This work was supported in part by the New York State Empire State Development Division of Science, Technology and Innovation (NYSTAR) under contract C160142.