Modeling Errors Analysis in Inverse Dynamics Approaches within a Task-Priority Framework

Abstract

Control algorithms of large Degrees Of Freedom (DOFs) robotic systems based on Inverse-Kinematic or Inverse-Dynamic approaches are two well known topics of research in control theory. In the last years those subjects never diminished their importance due to the progressively increase of the number of DOFs in robotic applications. Among the various possibilities, the design of tasks arranged in priorities is one of the prominent. Task-priority at the kinematic or dynamic level is witnessing an increase interest especially in the humanoid or floating-base (underwater or aerial) communities. This paper investigates the effects of modeling errors in control algorithms based on Inverse Dynamics with respect to the uncertain knowledge of the dynamic parameters of the model. In particular, it is of interest to understand the effect on the null-space projections and the sources of steady state errors.