Cable Manipulation with a Tactile-Reactive Gripper (1910.02860v3)

Abstract: Cables are complex, high dimensional, and dynamic objects. Standard approaches to manipulate them often rely on conservative strategies that involve long series of very slow and incremental deformations, or various mechanical fixtures such as clamps, pins or rings. We are interested in manipulating freely moving cables, in real time, with a pair of robotic grippers, and with no added mechanical constraints. The main contribution of this paper is a perception and control framework that moves in that direction, and uses real-time tactile feedback to accomplish the task of following a dangling cable. The approach relies on a vision-based tactile sensor, GelSight, that estimates the pose of the cable in the grip, and the friction forces during cable sliding. We achieve the behavior by combining two tactile-based controllers: 1) Cable grip controller, where a PD controller combined with a leaky integrator regulates the gripping force to maintain the frictional sliding forces close to a suitable value; and 2) Cable pose controller, where an LQR controller based on a learned linear model of the cable sliding dynamics keeps the cable centered and aligned on the fingertips to prevent the cable from falling from the grip. This behavior is possible by a reactive gripper fitted with GelSight-based high-resolution tactile sensors. The robot can follow one meter of cable in random configurations within 2-3 hand regrasps, adapting to cables of different materials and thicknesses. We demonstrate a robot grasping a headphone cable, sliding the fingers to the jack connector, and inserting it. To the best of our knowledge, this is the first implementation of real-time cable following without the aid of mechanical fixtures.

• The paper presents a novel tactile feedback framework for real-time cable manipulation without relying on external constraints.
• It employs a GelSight sensor together with PD and LQR controllers to precisely regulate gripping force and cable pose.
• Experimental results demonstrate adaptive cable following in diverse conditions, outperforming traditional open-loop methods.

Cable Manipulation with a Tactile-Reactive Gripper

The paper "Cable Manipulation with a Tactile-Reactive Gripper" presents a novel approach for the real-time manipulation of cables using a pair of robotic grippers equipped with high-resolution tactile sensors. The authors delineate a comprehensive perception and control framework capitalizing on tactile feedback to execute the task of cable following without external mechanical constraints.

Traditional methods often rely on conservative techniques, making use of mechanical constraints to simplify the manipulation of deformable objects, such as cables, which inherently possess high dimensionality and dynamic complexity. This research, however, takes a significant step forward by proposing a methodology that employs tactile feedback for direct cable interaction, facilitated by a dual-arm robotic setup. The paper explores the practical manipulation of freely moving cables using tactile sensing rather than visual or mechanical aids.

System Architecture and Methodology

The central innovation lies in the use of a vision-based tactile sensor, GelSight, which estimates the cable's pose within the grip and measures the friction forces experienced during cable sliding. This sensor provides a rich stream of tactile data, critical for the control framework. The hardware includes a lightweight, reactive gripper architecture capable of responsive actuation, which the authors designed to support real-time adjustments in response to the tactile feedback.

The perception and control framework comprises two main components:

1. Cable Grip Controller: This component utilizes a Proportional-Derivative (PD) controller with a leaky integrator to modulate the gripping force. The aim is to maintain suitable frictional sliding forces as the cable moves, ensuring effective grip.
2. Cable Pose Controller: An optimal Linear Quadratic Regulator (LQR) controller is deployed, which leverages a learned linear model of the sliding cable's dynamics. This controller maintains the cable's alignment and positioning on the fingertips.

Experimental Results

The research demonstrates that the proposed system is capable of following one meter of cable in various configurations, managing to adapt autonomously to cables of different thicknesses and materials. The experimental results indicate that the system can complete intricate tasks such as grasping a headphone cable, locating the jack connector, and inserting it.

The tests reveal the system's proficiency with various cables, assessed through metrics such as the length of cable correctly followed relative to its total length, normalized distance per regrasp, and normalized velocity. The implementation not only validates the system's capability for real-time cable following but also illustrates significant improvements when benchmarked against simpler, open-loop methods.

Implications and Future Work

The implications of this work are profound both theoretically and practically. Theoretically, the research contributes to the understanding of how tactile feedback can be integrated into the control loop of robotic systems handling deformable linear objects. Practically, it presents a roadmap for developing robots capable of complex manipulation tasks without needing heavy reliance on vision systems or fixed external constraints.

For future developments, potential enhancements could involve increasing the tactile sensor's feedback frequency and optimizing the finger-sensor geometry to improve manipulation of more complex cable geometries. Furthermore, leveraging deeper learning techniques or model-based reinforcement learning could enhance the robustness and adaptability of the control paradigms to newer scenarios of cable manipulation tasks.

This paper lays the groundwork for future explorations into tactile-based manipulation, opening new pathways in the realms of robotic dexterity and tactile sensory applications in handling deformable materials.