I made this website to present a part of my work and myself.
All that is presented here is in free access so anyone can get interest in, or used what i have done. Enjoy;
I am Vincent Samy, a current postdoc at the JRL1, Tsukuba, Japan.
During my PhD at LIRMM2, I worked on safe falls for humanoid robots, or 'how to control the robot so that it does not smack the ground too hard'.
As a fall is a relatively fast event (<1s), i have made a point of honor to develop on-line algorithms.
The main contribution of my thesis is, without doubt, the adaptive qp that enables to tune the motor's PD-gains on-line in order to comply to the impact.
I also used polytopes coupling with an MPC to predict forces the adaptive-qp should aim for.
I changed field and I currently work on musculoskeletal models.
Through my different experiences, i have been working on diverse subject.
From the manipulability of hybrid actuator to humanoids control and more specifically, safe falls. I have worked in laboratories like the complex mechanical system engineering laboratory at Kobe university2, Japan, the LIST3 at CEA4, France, ISIR5 of the Pierre et Marie Curry university, France, and the Lirmm.
Now, I mostly implement libraries for researches on musculoskeletal models.
Through the years i have developed different tool to help me for my work but also for fun.
I drop on my github repository all libraries I can upload freely.