Dynamically stable trajectory synthesis for a biped robot during the single-support phase

The aim of this paper is to develop an algorithm to generate dynamically stable walking trajectories for a biped robot during the single-support phase, which can be used as reference trajectories for control. The biped is modeled as an anthropomorphic planar mechanical system of 7 links connected by 6 actuated rotoidal joints. We want the biped robot to move with a human-like behavior. In order to achieve this goal the kinematics and dynamics aspects have been treated separately. To obtain a human-like motion the biped has been regarded as a redundant manipulator, under the assumption that the supporting foot was fixed to the floor. In particular, the concept of task priority in relation to the inverse kinematics problem has been exploited. To ensure that planned motion fulfils the dynamics requirements too, the attention has been focused on the interaction between stance foot and ground, which has been carefully studied by means of classical mathematical tools such as the center of pressure point (or ZMP). A detailed simulation of biped trajectory is presented to show the effectiveness of the proposed approach.