Organisation, Repräsentation und Analyse menschlicher Ganzkörperbewegung für die datengetriebene Bewegungsgenerierung bei humanoiden Robotern
- KIT Scientific Publishing 2017
- 1 electronic resource (VIII, 308 p. p.)
Open Access
This work presents an approach to data-driven motion generation for humanoid robots, which is based on the observation and analysis of human whole-body motions. To this end, we investigate how captured human motions can be represented, classified and organized in a large-scale motion database. The statistical modeling of the transitions between characteristic whole-body poses enables the subsequent generation of multi-contact motions.