Humanoid Gait Synthesis Using Trajectory Plot and Relative-ZMP (R-ZMP) Concept

Abstract

The human body exploits the redundancy of the degree of freedom (DOF) to execute various motions and maintain the body stability. In Humanoid Robots (HRs), the Zero Moment Point (ZMP) and Center of Gravity (COG) are commonly used to evaluate stability. One of the basic motions of a HR is walking. There are two types of walking scheme, namely the Static and Dynamic Gait. The Static Gait uses the COG as the stability criterion. Static Gait usually yields a very slow walking gait. Hence, static walk is now not very common in HRs. On the contrary, Dynamic Gait employs the ZMP as the criterion of stability. Robots like ASIMO and QRIO use Dynamic Gait scheme to achieve impressive walking speed and stability. In this paper, a new trajectory design algorithm by using the trajectory plot and posture plot is proposed. This method is based on the new Relative-ZMP (R-ZMP) concept. The resulting walking gaits were then tested on a humanoid with 18 DOF, LUCY. It is evident that it is able to walk much faster in comparison with other humanoid robots which participated in the Humanoid Robocup Competition 2006.