Optimization of the Intrinsic Stiffnesses of a Robot with Variable-Stiffness-Actuators under External Disturbances with Respect to Control Effort

Abstract

Variable stiffness actuators (VSA) are used in robotics especially for the safe interaction with humans and because of their energy efficiency when it comesto cyclic movements. The elastic structure preserving impedance (ESPi) control framework is a control concept for VSAs. So far it is unclear how the intrinsic stiffness setting of the VSA is supposed to be set when the robot is subjected to external disturbances while controlled by the ESPi controller. The main goal of this thesis is to utilize the variable stiffness of VSAs in order to minimize the control effort and the power input of the motor during external disturbance. The focus lies on sinusoidal disturbances with single and multiple frequencies. To analyze the impact of the intrinsic stiffness setting on the control effort and the power input, four different optimization problems are formulated and solved to calculate optimal intrinsic stiffness settings. They cover both linear and nonlinear VSAs. This work shows the potential of the minimization of the control effort by calculating optimal intrinsic stiffness values. Performing the minimization of the power input of the motor did not prove to have a high impact since the influence of the stiffness setting on the power input is relatively low. It remains to be investigated whether it is possible to calculate the optimal intrinsic stiffness values in real time.