Evaluating the Use of Support Vector Machine Models for Contact State Identification of Non-Rigid Cylindrical Parts During Robotic Assembly

The article solves the problem of determining the current contact state (CS) during assembly using the support vector machine (SVM) method when receiving information about the position coordinates by the industrial robot system and forces from a force-torque sensor. The purpose of the work is to create a method for recognizing the contact state when assembling non-rigid cylindrical parts with rigid sleeve as a basis for developing an algorithm for the coupling process. A mathematical model of the problem of recognizing the contact state when coupling non-rigid cylindrical peg-in-hole type parts by a robot retrofitted with a force-torque sensor is presented. In this article, it is proposed to use the SVM method with a data source that is the coordinates of the position from the industrial robot system and the forces from the force-torque sensor to identify the contact state of compatible robot movement tasks. An experimental setup with an ABB IRB-140 industrial robot equipped with a Schunk force-torque sensor is used to evaluate the effectiveness of the proposed method.