GAIT GENERATION FOR THE HUMANOID NAO BASED ON HUMAN WALKING PARAMETERIZATION

The work presented here focus on humanoid walking. More precisely we deliberately divide the control architecture in two distinct parts, coordination and stability. The goal is to design robot coordination mimicking the human one, by reducing the latter's complexity to a minimal set of relations. In our previous work, polynomial approximations were done and implemented into a kinematic simulation of an anthropomorphic humanoid model. This simulation showed smooth biomimetic walking at different speeds. This paper presents the real-world implementation of our methodology using the humanoid robot NAO. The results are promising for applying this entire work on more complex humanoid robots, other gaits, and gait transitions.