Fuzzy variable stiffness in landing phase for jumping robot

Abstract

Some important applications of humanoid robots in the nearest future are elder care, search and rescue of human victims in disaster zones and human machine interaction. Humanoid robots require a variety of motions and appropriate control strategies to accomplish those applications. This work focuses on vertical jump movements with soft landing. The principal objective is to perform soft contact allowing the displacement of the Center of Mass (CoM) in the landing phase. This is achieved by affecting the nominal value of the constant parameter P in the PID controller of the knee and ankle motors. During the vertical jump phases, computed torque control is applied. Additionally, in the landing phase, a fuzzy system is used to compute a suitable value for P, allowing the robot to reduce the impact through CoM displacement. The strategy is executed on a gait robot of three Degrees of Freedom (DoF). The effect of the impact reduction is estimated with the calculations of the CoM displacement and the impact force average during the landing phase.