Towards Automated Connector Assembly: Wire Insertion Combining Tactile and Vision Sensors

The manipulation of flexible objects, such as cables and wires, remains a challenge for automation in industries like automotive, aerospace, and electronics, where high precision and dexterity are required. A critical task is assembling connectors, which involves accurately inserting a wire terminal (pin) into a connector hole. Despite its importance, this task is still predominantly performed manually due to the complexity of handling flexible objects and high accuracy demands. This paper presents a novel methodology for automating connector assembly using a robotic arm combined with a mechatronic tool: a parallel-jaw gripper with sensorized fingers and an RGB camera. The sensorized fingers, equipped with tactile sensors, estimate the local shape and position of the wire, while the camera determines the axial orientation of the pin for alignment. Tactile sensors are also exploited during insertion, where a feature indicator is utilized to ensure correct task execution. A post-insertion pull test validates proper locking. Data-driven techniques applied to tactile and vision data improve accuracy and ensure precise execution during assembly. The proposed approach is experimentally validated on a real robotic system, demonstrating its feasibility and effectiveness as an important step toward fully autonomous connector assembly tasks.