Having recently returned from a research visit in Plymouth University, I had to prepare a report for EUCog Network that sponsored this event. While wrapping up all the exciting research that was fiercely taking place in what seems to have been an amazing summer for northern Europe, I came to realize the deep aims of EUCog.


EUCogIII ROBOSKIN Default

I have recently been informed that the EUCogIII Coordination Council has voted in favour of supporting in full my research summer visit at the Centre for Robotics and Neural Systems in Plymouth University.



A Nao equipped with a tactile sensitive fingertip performing a stroke on another tactile sensitive NaoGreat news indeed! I will be spending the summer in Plymouth University at the Centre for Robotics and Neural Systems. In what seems to be an amazing collaboration built on the ROBOSKIN project you will be able to track my progress through this page.

This research visit is intended on performing experiments for learning tactile gestures on real humanoid robots using a novel hierarchy of Self-Organizing maps that has been developed during my PhD.

In the past, similar hierarchical models were introduced by developmental psychologists; however, these computational models have only been demonstrated in abstract domains to relate their learning features to infant cognition, e.g., understanding of causal events, or for gesture recognition. By extending these models, we have made it possible to learn motor skills via imitation in a real humanoid robot.


During this visit, I will have the opportunity to explore social motor skills in tactile sensitive robots and whether it is possible to further extend it in the Reinforcement Learning domain. An extensive analysis will not only help me support the contributions of my PhD, but will also establish the applicability of these algorithms on a field as complex as humanoid control.



Following this week's puzzle in New York Times wordplay blog I found out about Dr. Nelson Blachmanan amazing mathematician. Producing mathematical research is difficult, let alone for Dr. Blachman who is 90 years old this year and also blind, however, he is still an active mathematician, an expert user of Mathematica and uses computers without a glitch (video).

Yesterday, just after 2:00 am, almost falling asleep, I went through my twitter feed for a last time before making a move. It was not long before Steven Strogatz, an inspiring mathematician, writer, and Professor in Cornell (it is worth following him on twitter @stevenstrogatz), tweeted about his solution on the "Theater problem" also featured on this NYTimes blog article.  The complete chapter "All About e" of his book "The Joy of x" has been released by the publisher "Houghton Mifflin Harcourt" (Richard Stallman would have said here, we support the authors not the publishers!).

It's been quite a lot time by now that I have developed a tiny library to read files from the CMU MoCap lib. In case you are working with the .amc files you may find it useful.

I was inspired by the Matlab version of the interface, however, I have not implemented the writing back to .amc file functionality. At this version, you can only read the motion files. The examples should be self-declarative and comprehensive.

You can find or clone the library at:

http://bitbucket.org/gpierris/pycmu_mocap or    git clone This email address is being protected from spambots. You need JavaScript enabled to view it.:gpierris/pycmu_mocap.git


In the last year, or even more, we have put all our efforts into the development of a novel Reinforcement Learning algorithm for temporal pattern learning. In our approach, we use a Hierarchy of self-organizing maps to encode sequences of motion primitives. More details of this work are available in the publications section and as soon as we finish the review more work will be out.

To make long story short...

During the National Eisteddfod Event, Ebbw Vale, Wales, UK, August 6-7, 2010, I developed an interface for the Animazoo Gypsy-6 full-body motion capture to control a simulated Aldebaran Nao robot. We had the opportunity to try it on the famous football player Malcolm Allen, where he managed to kick a couple of penalties!

Another project to control the Aldebaran Nao humanoid robot using the Kinect Sensor has been put in the back burner for now.