The iCub humanoid robot is an open-system, meaning both open-source and open-hardware, robotic platform developed in EU funded FP7 research projects. It provides (in our setup) a 41 degree-of-freedom (DOF) upper-body, comprising two arms, a head and a torso. The iCub is generally considered an interesting experimental platform for cognitive and sensorimotor development and embodied Artificial Intelligence research, with a focus on object manipulation.
Juxi LeitnerResearcher Robotics & AI
iCub Research & Projects
Papers
Recent iCub Related Research Publications:
Reactive Reaching and Grasping on a Humanoid: Towards Closing the Action-Perception Loop... 2014
Improving Robot Vision Models for Object Detection Through Interaction. 2014
Curiosity driven reinforcement learning for motion planning on humanoids. 2014
Artificial Neural Networks For Spatial Perception: Towards Visual Object Localisation in Humanoids. 2013
Task-Relevant Roadmaps: A Framework for Humanoid Motion Planning. 2013
Learning Visual Object Detection and Localisation Using icVision., 2013
Learning Spatial Object Localization from Vision on a Humanoid Robot, 2012
Autonomous Learning Of Robust Visual Object Detection And Identification On A Humanoid, 2012
Transferring Spatial Perception Between Robots Operating In A Shared Workspace, 2012
Research & Outcomes
The overarching goal of my iCub related research is to create robots with more autonomous and more adaptive behaviours, leading to more `intelligent´ robots. My research is focussed at the intersection between IDSIA's Machine Learning (Reinforcement Learning) group and robotics. My goal was to make the iCub see, that is, developing computer vision algorithms for object detection, identification and localisation.
Software Frameworks
Recently icVision, my modular computer vision system used in our cognitive robotics research, was released. Have a look at the project page if you are interested. It is an easy-to-use, modular framework performing vision related tasks on the iCub humanoid robot for research experiments.
LEOGrasper is now available as well. It is our light-weight, easy-to-use, one-shot grasping system for the iCub. It's been used extensively at IDSIA, especially for the IM-CLeVeR video.
Perception and Sensorimotor Coordination
We submitted a paper to the WCCI 2014 (IJCNN Special Session on Robot Vision) on our research about choosing actions to improve object detection in robotic vision. The video is here:
Intrinsically Motivated Agents (IM-CLeVeR)
My PhD is partly funded by the EU FP7 project IM-CLeVeR, which includes partners from CNR (in Rome), Universities of Ulster and Aberstwyth (UK). This project focusses on the iCub robot and how it can learn to interact with its environment. At IDSIA we put a lot of work into making the iCub learn how to perform object manipulation.
Tele-operation of a Complex Humanoid
The operation of complicated advanced robot systems is not a trivial task. How an operator can perform certain tasks intuitively and without breaking the robot
We worked on this and are now able to control the iCub with a variety of input devices, using e.g. a LEAP motino device, a six-axis joy-pad or even a simple accelerometer (such as the one in your phone)!
More cool things
Stereo-Vision Depth Perception in Humanoid Robots
This project in 2007 at IST in Lisbon aimed to give an the experimental evaluation of a dense disparity estimation algorithm, focusing on the most relevant aspects for humanoid robots: real-time functionality and ability to deal with calibration errors. The method and its real time implementation (using C++ and open source libraries) are tested in a variety of cases to show its performance and the quality of its output as a function of the design parameters. An early version of the iCub hardware (mounted on-top a torso with one robotic arm) together with the iCub abstractions for the hardware (YARP) were used.
Shakey Award Winner - AAAI 2013 Video
Our video showing the iCub doing cool motions was awarded Best Student Video at the AAAI Video Competition!!