Global Sensitivity Analysis of Dynamic Parameters to Industrial Robots Considering Spacetime Two-Dimensional Output Responses

Abstract

In practical engineering, the joint torques of a 6-degree-of-freedom (6-DoF) industrial robot can be measured in two dimensions of space and time, i.e., the torque response of six joints in different spatial positions at a certain moment and the torque response of one joint at multiple moments. Accurate quantification of the sensitivity of dynamic parameters considering spacetime two-dimensional output responses of joint torque is of great significance for improving the accuracy and efficiency of dynamic parameter identification. Therefore, we implement the variance-based global sensitivity analysis (GSA) method to quantify the sensitivity of dynamic parameters from two dimensions of space and time. First, a linearized dynamics model of the industrial robot is established and the optimal excitation trajectories are used to stimulate dynamic parameter characteristics. Then, some dynamic parameters are selected as uncertain input variables for GSA. At last, the influence of dynamic parameters on the torque output responses of the dynamics model in two dimensions of space and time is quantified effectively by GSA. In this study, results show that GSA can provide more adequate information about the sensitivity of dynamic parameters of industrial robots. And GSA is helpful for dynamic parameter recombination and experimental data selection when identifying the dynamic parameters of industrial robots. In addition, this article also illustrates that the sensitivity of dynamic parameters directly reflects whether the excitation trajectory can fully stimulate the dynamic characteristics of industrial robots, and provides guidance for the determination of the excitation trajectory in the dynamic parameter identification process.