Resource Menu


Robbie is an experimental robot based on Keridwen and Java/OSGi approach for the software layer and a Raspeberry Pi and ARM hardware configuration. Robbie is a simple proof of concept to evaluate the relevance of the Keridwen Java / OSGi based approach for very light and low cost hardware infrasctructures, like embedded systems, robots, low-cost scientific measurement and monitoring system.

Robbie includes its own exploration algorithms to perform its missions in an autonomous manner as much as possible and a Web based remote control interface. An image analysis system based on OpenCV and JavaCV allows Robbie to perceive its environment. A double GPS and radio and IR beacons based localisation systems, linked to a light GeoTools based IGS, are integrated in order to localised the rover in various contexts. For ground applications and to siplify the robot/human interaction, a Text-to-Speech (TTS) and voice controles services complete the user interface.

Thanks to the Pi4j project, all devices are controlled from the Raspberry Pi through the Java layer.

Robbie is de facto a modular platform to:

  • Explore modular software design in robotics for a better extensibility and adaptability to taylored applications;
  • Through a realistic application:
    • Valid the capability of Keridwen for modular, services oriented applications;
    • Evaluate the constraint related to a remote control and data transfer on low band-width and high latency system;
    • Explore scientific and operational data capture and remote processing;
  • Explore robot-environment interaction algorithms, to facilitate its autonomy and survival;
  • Explore the human-robot interactions;
  • Through 3D simulation done with the SPIS and the GEANT-4 based software, evaluate the capabilities of low-cost hardware components in space environments, especially in lander-plasma interactions and radiations effects analysis.
For further information, please contact us:

Embbeded technologies

Get involved (login required)

Last edited by Julien Forest at Aug 22, 2017 2:10 PM - Edit content - View history - View source