Robust Task-Space Quadratic Programming for Kinematic-Controlled Robots

Task-space quadratic programming (QP) is an elegant approach for controlling robots subject to constraints. Yet, in the case of kinematic-controlled (i.e., high-gain position or velocity) robots, closed-loop QP control scheme can be prone instability depending on how gains related tasks constraints are chosen. In this article, we address such shortcomings. First, highlight nonrobustness system against nonmodeled dynamics, as those relative joint flexibilities, external perturbations, etc. Then, propose a robust formulation based high-level integral feedback terms task space including The proposed method formally proved ensure stability and intended applied any under practical assumptions. We assess our through experiments fixed-base robot performing stable fast motions floating-base humanoid robustly reacting perturbations keep its balance.