Tactile Skewness Analysis and Contact State Recognition by Fusing Tactile and Wrist Force for Flexible Printed Circuits Assembly Task

Abstract

With the rapid progress of the computer, communication, and consumer (3C) industry, robotic assembly tasks have become more important and challenging, such as the assembly of flexible printed circuits (FPCs) in mobile phones. Due to the interference fit and the low stiffness of the connection, the assembly of FPC faces two primary challenges: precise positioning as well as compliance and safety of contact. Vision-based methods cannot sense contact forces, which may result in damage, while six-dimensional force methods fail to adequately capture the contact state between connectors. In contrast, tactile feedback can resolve these challenges. Current tactile-based FPC assembly strategies only detect the FPC’s position, overlooking the mobile phone’s position error and the error introduced by the transfer after identification. This article proposes a novel FPC assembly strategy that fuses tactile feedback and six-dimensional force to address existing challenges. First, a tactile skewness analysis model is proposed to correct the end-effector’s orientation error. Next, a contact state recognition model is developed by combining the tactile data from the contact surface and the six-dimensional force from the robot arm when the male and female connectors are in contact, which enables the classification of position error directions. The position error is then corrected through the proposed directional searching and snap-fitting strategy. Finally, experiments validate the proposed method, demonstrating its effectiveness in reducing both orientation and position errors. Compared to traditional methods, it improves assembly efficiency and prevents snap-fitting failure caused by male and female connectors losing contact.