6,090 research outputs found
INTELLIMAN. WP5 T5-3-1. Attention-Based Cloth Manipulation from Model-free Topological Representation
No full textThis dataset contains data related to a novel attention-based neural architecture capable of solving a smoothing task for deformable objects, such as clothes and fabric, by means of a single robotic arm. In particular, the dataset contains the data used for the training on a model-free learning-based policy for cloth smoothing task and the script used to collect the resulting data. The data were produced in the framework of Horizon Europe INTELLIMAN project and are presented in the publication:
K. Galassi, B. Wu, J. Perez, G. Palli and J. -M. Renders, "Attention-Based Cloth Manipulation from Model-free Topological Representation", 2024 IEEE International Conference on Robotics and Automation (ICRA), Yokohama, Japan, 2024, pp. 18207-18213, doi: 10.1109/ICRA57147.2024.10610241
REMODEL. WP5. T5_5_2. Cable Detection and Manipulation fo DLO-in-Hole Assembly Task
No full textThe dataset contains the pull test executed after the DLO executed in the paper:
K. Galassi, A. Caporali and G. Palli, "Cable Detection and Manipulation for DLO-in-Hole Assembly Tasks," 2022 IEEE 5th International Conference on Industrial Cyber-Physical Systems (ICPS), 2022, pp. 01-06, doi: 10.1109/ICPS51978.2022.9817006.
The test where conducted by inserting the DLO in an switchgear component after correcting the position using the proposed algorithm. After the screw where tighted, the robot verify the correction of the experiment by pulling the DLO and collecting the data. The data registered show the forces and torque along the built-in robot force sensors
REMODEL. WP5 T5-4-2. Robotic Wires manipulation for Switchgear Cabling and Wiring Harness Manufacturing
No full textThe dataset contains the data validating a cyber-physical system for cable manipulation composed by a robotic arm, a parallel industrial gripper and a pair of tactile sensors. The data were obtained conducting multiple tests in an experimental setup in which a cable must be routed along two linear paths connected by a turn and with four fixing points. Specifically, the tests consisted in the analysis and the evaluation of the proposed PID tensioning controller alongside the active tension control of the gripper implemented in the research for the manipulation of deformable linear object for switchgear manufacturing. The data are gathered using ROS and saved in a .bag file, the image proposed in the publication can be obtained with the matlab file attached to the dataset. The data are presented in the publication:
K. Galassi and G. Palli, "Robotic Wires Manipulation for Switchgear Cabling and Wiring Harness Manufacturing," 2021 4th IEEE International Conference on Industrial Cyber-Physical Systems (ICPS), 2021, pp. 531-536, doi: 10.1109/ICPS49255.2021.9468128
INTELLIMAN. WP5 T5-1-1. Monocular Estimation of Connector Orientation: Combining Deformable Linear Object Priors and Smooth Angle Classification
No full textThe dataset contains data related to the development of a learning-based solution for the detection of manufacturing errors in the context of robotic wiring harness assembly. In particular, the dataset contains the training set, the validation set, and the test set used to develop the proposed neural networks-based error detector, and the scripts to train and test the neural networks and store the trained models. The data were produced in the framework of Horizon Europe INTELLIMAN project and are presented in the publication:
A. Caporali, K. Galassi, G. Berselli and G. Palli, "Monocular Estimation of Connector Orientation: Combining Deformable Linear Object Priors and Smooth Angle Classification," 2024 IEEE International Conference on Advanced Intelligent Mechatronics (AIM), Boston, MA, USA, 2024, pp. 799-804, doi: 10.1109/AIM55361.2024.10637081
INTELLIMAN. WP5 T5-1-2. Scalable Shared Encoding Architecture for Learning-Based Error Detection in Robotic Wiring Harness Assembly
No full textThe dataset contains data related to the development of a learning-based solution for the detection of manufacturing errors in the context of robotic wiring harness assembly. In particular, the dataset contains the training data, the validation data, the needed python script to load the data, and the model used for the training of the neural network, used to develop the proposed neural networks-based error detector. The data were produced in the framework of Horizon Europe INTELLIMAN project and are presented in the publication:
Galassi, A. Caporali, G. Laudante and G. Palli, "Scalable Shared Encoding Architecture for Learning-Based Error Detection in Robotic Wiring Harness Assembly", 2024 IEEE International Conference on Advanced Intelligent Mechatronics (AIM), Boston, MA, USA, 2024, pp. 518-523, doi: 10.1109/AIM55361.2024.10637054
REMODEL. WP5. T5_2_3. Combining Vision and Tactile Data for Cable Grasping
No full textThe dataset contains the visual results obtained from deformable linear objects (DLOs) grasping experiments performed in the framework of REMODEL project. These experiments were focused on properly combine vision and tactile data to locate a DLO and grasp it according to a required position and orientation. The robot is programmed to grasp a wire and bring it in front of a camera (intrinsic and extrinsic parameter of the camera needs to be known), then a picture is taken and using Ariadne+ is obtained the position of the wire not occluded by the gripper while the remaining part reconstructed by the tactile sensor. In the dataset are also included: (1) the data reading from the tactile sensor developed inside the REMODEL project provided by UCLV and used for these experiments; (2) the corresponding code used to reproduce the results (the executable is compatible with all the ROS’s compatible robot with the tactile sensor). More specific information about the method and the results can be found in the paper: A. Caporali, K. Galassi, G. Laudante, G. Palli and S. Pirozzi, "Combining Vision and Tactile Data for Cable Grasping", 2021 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), 2021, pp. 436-441, doi: 10.1109/AIM46487.2021.951744
REMODEL. WP5. T5_5_3. ROSAPP for Deformable Objects Grasping and Shape Detection with Tactile Fingers and Industrial Grippers
No full textThe dataset presented by UNIBO consist in two different tests aimed to evaluate the tactile sensor. The experiment consist in the grasping on a DLO (electrical cable) and the correction of robot position using the information from the tactile sensor. In the ‘linear’ experiment the robot move along z. In the ‘angular’ experiment, the robot move along z and correct the orientation of the gripper. In the data are gathered the result, showing the robot position and grasped cable position first order estimation on the sensor before and after the correction
Robotic Learning for Perception and Manipulation of Deformable Objects
No full textL'abstract è presente nell'allegato / the abstract is in the attachmen
Robotic wires manipulation for switchgear cabling and wiring harness manufacturing
No full textThis paper describes a cyber-physical system for a wiring operations, composed by a robotic manipulator and a gripper with tactile sensors. This system can be used for switchgear cabling and manufacturing of wiring harnesses. The manipulation of electrical wires, and more in general deformable linear objects, is a complex task of large interest for different industrial applications. The proposed system is designed to shape a cable along a desired path including fixing points and obstacles exploited to shape the cable itself. To accomplish this goal, the wire position needs to be determined starting from a list of point and then converted in a joint reference pose for the manipulator to reproduce the desired trajectory. Moreover, the wire tension must be controlled on the basis of both the estimation of the robot external wrench and the tactile data by acting on the gripper finger opening. An experimental setup in which a cable must be routed along two linear paths connected by a turn and with four fixing points has been used to validate the proposed solution
- …
