Scribd Logo
Upload

Welcome to Scribd!

  • 25 views

    Trabajo Abraham Leon

    Uploaded by

    Darío Acosta
    This paper proposes a method to estimate the interaction force between a robot and objects during grasping and picking using only vision, without tactile or force sensors. A deep neural network is designed that combines DenseNet and a transformer to predict interaction force from images. The network is trained on a generated database of robot interactions with 41 objects. The proposed method is able to estimate interaction forces for both known and new objects, and outperforms other models with less than 3% error. The results demonstrate that visual information alone can be used to intelligently estimate robot grip forces during interaction.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Trabajo Abraham Leon

    Uploaded by

    Darío Acosta
    25 views13 pages
    This paper proposes a method to estimate the interaction force between a robot and objects during grasping and picking using only vision, without tactile or force sensors. A deep neural network is designed that combines DenseNet and a transformer to predict interaction force from images. The network is trained on a generated database of robot interactions with 41 objects. The proposed method is able to estimate interaction forces for both known and new objects, and outperforms other models with less than 3% error. The results demonstrate that visual information alone can be used to intelligently estimate robot grip forces during interaction.

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This paper proposes a method to estimate the interaction force between a robot and objects during grasping and picking using only vision, without tactile or force sensors. A deep neural network is designed that combines DenseNet and a transformer to predict interaction force from images. The network is trained on a generated database of robot interactions with 41 objects. The proposed method is able to estimate interaction forces for both known and new objects, and outperforms other models with less than 3% error. The results demonstrate that visual information alone can be used to intelligently estimate robot grip forces during interaction.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    25 views13 pages

    Trabajo Abraham Leon

    Uploaded by

    Darío Acosta
    This paper proposes a method to estimate the interaction force between a robot and objects during grasping and picking using only vision, without tactile or force sensors. A deep neural network is designed that combines DenseNet and a transformer to predict interaction force from images. The network is trained on a generated database of robot interactions with 41 objects. The proposed method is able to estimate interaction forces for both known and new objects, and outperforms other models with less than 3% error. The results demonstrate that visual information alone can be used to intelligently estimate robot grip forces during interaction.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 13

    LINEAMIENTOS PARA TRABAJOS ESCRITOS

    PORTADA

    Deberá ser la primera hoja a la vista y contendrá la siguiente información:


    S.E.P. S.E.S.
    TecNM

    Instituto Tecnológico de Aguascalientes

    Número de Control: 22150 619

    Maestro: Edgar Darío Acosta Pérez

    Especialidad: Electrónica
    Alumno(a): Abraham Alejandro León Durón
    Materia FUNDAMENTOS DE INVESTIGACION
    A new link-legged insulator inspection robot

    Shangqun Dong

    Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China


    August 29, 2022

    Insulator stringsLink-legged robotChebyshev linkageInspectingClimbing

    The robot is designed as for a flat surface to be able to carry other execution mechanisms,
    such as cleaning or vision equipment. The objective of which is to make the robot climb and
    overcome the deficiency of a wide variation of speed in the direction of movement. They made
    a prototype of the robot and conducted an experiment in the laboratory, demonstrating that
    the robot can scale and inspect both vertical and horizontal insulator strings as desired.

    Due to increasing labor shortages and safety requirements in insulator string inspection,
    inspection robots have been a primary alternative to humans in such a harsh environment.
    This paper presents a newly designed knuckle-legged insulator inspection robot based on the
    Chebyshev link, which can inspect deteriorated insulators and climb both vertical and
    horizontal insulator strings.

    they don't show how they did it

    The link-legged insulator inspection robot is based on the Chebyshev linkage.

    The robot can climb both vertical and horizontal insulator strings.

    The robot can carry other executing mechanisms, such as cleaning equipment.

    The design is verified in our laboratory by a prototype.


    Remote control of a robotic prosthesis arm with
    six-degree-of-freedom for ultrasonic scanning and
    three-dimensional imaging

    Qinghua Huang

    School of Electronic and Information Engineering, South China University of Technology,


    Guangzhou, 510006, China

    30 de julio de 2019

    Robot-assisted ultrasound examination Teleoperated medical imaging 3D ultrasound imaging


    Human-machine interaction Multi-degree-of-freedom robotic prosthesis

    Los sistemas de ultrasonido robóticos se han convertido en uso clínico en las últimas
    décadas, aumentando la precisión y la calidad de las operaciones médicas. En este
    artículo, proponemos un sistema de escaneo de ultrasonido asistido por robot y
    telecontrolado para imágenes tridimensionales.

    A six-degree-of-freedom (DOF) robotic arm acting as an upper-limb prosthesis remotely


    controlled by an operator drove the ultrasound probe to scan skin surfaces. Via the Internet,
    the operator at a distant location monitored the robotic prosthesis and the patient through
    four on-site video cameras

    Based on the real-time video transmissions, the operator could manipulate a joystick to send
    instructions from the remote operation terminal to the system at the patient's side to adjust
    the movements of the probe. In this way, the safety and accuracy of the proposed system
    could be guaranteed. In vitro and in vivo experiments were performed to validate the viability
    and performance of the proposed system. Quantitative experimental results show that the
    volume measurement error was less than 1.1%, indicating that our system can precisely and
    flexibly control the probe scan and produce high-quality 3D ultrasound images.
    is the graphs show the results of the graphs of the results

    Real-time audio and video streaming is also done on our system. The system can be used with
    the wired network, 4G and 5G wireless network. The system was tested on 95 subjects in
    clinical trials
    Robotic index finger prosthesis using stackable
    double 4-BAR mechanisms

    Giho Jang

    Department of Electronic Systems Engineering, Hanyang University, Ansan 426-791, South


    Korea

    July 30, 2019

    4 bar double stacking mechanism, Electromyography (EMG), Robotic Index Finger Prosthesis

    A robotic index finger prosthesis realized with one degree of freedom by using double 4-bar
    stackable mechanisms. In addition, their monitoring method uses two electromyographic
    (EMG) signals measured on the skin surfaces of the flexor digitorum superficialis (FDS) and
    extensor indicis (EI) in the lower arm.The measured raw EMG signals are processed using
    sequential procedures, such as root mean square, thresholding to extract the initial burst part,
    subtraction of the antagonistic EMG signal, and integration per 2 milliseconds. Finally, the
    effectiveness of the suggested mechanism and control method is verified through
    experiments.they do not show how they do
    Human–robot collaboration and machine learning:
    A systematic review of recent research

    Francesco Semeraro

    Cognitive Robotics Laboratory, The University of Manchester, Manchester, United Kingdom

    10 August 2022

    Human–robot collaboration,Collaborative robotics, Human–robot


    interaction,Machine learning
    Technological progress increasingly envisions the use of robots
    interacting with people in everyday life.Human-robot collaboration
    (HRC) is the approach that explores the interaction between a human
    and a robot, during the realization of a common goal, at a cognitive and
    physical level. In HRC works, a cognitive model is usually built, which
    collects inputs from the environment and the user, elaborates them
    and translates them into information that can be used by the robot
    itself.
    Summary

    Technological progress increasingly contemplates the use of robots that interact with people
    in everyday life. Human-robot collaboration (HRC) is the approach that explores the
    interaction between a human and a robot, during the realization of a common goal, at a
    cognitive and physical level. In HRC works, a cognitive model is usually built, which collects
    inputs from the environment and the user, elaborates them and translates them into
    information that can be used by the robot itself. Machine learning is a recent approach to
    build the cognitive model and behavioral block, with high potential in HRC. Consequently, this
    article proposes a comprehensive review of the literature on the use of machine learning
    techniques in the context of human-robot collaboration. 45 key articles were selected and
    analyzed and a grouping of works based on the type of collaborative tasks, evaluation metrics
    and modeled cognitive variables is proposed. Then, an in-depth analysis is carried out on
    different families of machine learning algorithms and their properties, along with the detection
    modalities used. Among the observations, the importance of machine learning algorithms to
    incorporate time dependencies stands out. The salient features of these papers are then
    cross-analyzed to show trends in HRC and provide guidelines for future work, comparing them
    to other aspects of HRC that do not appear in the review.
    The results are shown in the tables.
    Vision-based interaction force estimation for robot grip
    motion without tactile/force sensor
    Dae-KwanKo

    Force Estimation,Robot Grip,Interaction Force,Machine Learning


    10 August 2022

    Humans perceive an interaction force through the kinesthetic sense or tactile sense. By
    viewing the image, they estimate the force of interaction based on pseudo-haptics. The
    interaction force of a robot is traditionally measured using a contact-type tactile sensor or a
    force/torque (F/T) sensor. In this work, we propose a method to estimate the interaction force
    between a robot and objects during grasping and picking.We propose a deep neural network
    structure that combines DenseNet and a transformer encoder/decoder to predict the
    interaction strength. We verify the proposed network with the generated database that
    recorded the interaction with 41 objects. Additionally, we compare the results with the
    change of the network inputs. The model could estimate the interaction strength of various
    input modalities for both known and invisible objects during its training. The results clearly
    indicate that the proposed method produces the best results compared to other models, with
    less than 3% error in the estimation of the interaction force.

    Method for estimating the interaction force of robot without Force-Torque sensor.

    Robot grip force estimation with visual information.

    Intelligent integration of various input modalities.

    The proposed method could estimate the interaction force about seen/unseen objects.

    You might also like