Abstract

This dataset provides the source code used to implement and evaluate a robotic system for automated connector assembly involving deformable linear objects (DLOs). The proposed approach targets precision-critical connector insertion tasks, which are particularly challenging due to object flexibility and tight alignment tolerances. The stereo camera setup is used to estimate the full 6-D pose of connector pins, enabling accurate alignment prior to insertion, while the sensorized fingers monitor the insertion process in real time, reducing failure risks without relying on global force/torque measurements. The released material describe the implementation of the vision and tactile algorithms. The provided code and resources were developed and validated through real-world experiments on two distinct connector assembly tasks, as presented in the associated publication: A. Caporali, M. Mirto, S. Pirozzi and G. Palli, "Vision and Tactile Sensing for DLO Manipulation and Pin Insertion in Robotic Connector Assembly," in IEEE/ASME Transactions on Mechatronics, doi: 10.1109/TMECH.2026.3654272.