This paper proposes a method of humanoid robot data structure definition based on the characteristics of a graph with tree structure. It applies the depth-first search algorithm to traverse and uses screw theory and chain of multiplication to describe the robot's postures. The platform references zero moment point theory as the stability constraint of the robot's dynamic walking and plans the gaits accordingly. It also realizes the visualization of motion control. This study provides superior means and methods to the evolution of a humanoid robot.
展开▼