Skip to main content

ESR12 - Grasping and Manipulation

Implementation of spiking tactile system on a 7-DoF manipulator and development of manipulation algorithms for stable grasping of objects of different shape and softness.

TAGS   Biological Touch   Prosthetics   Technologies for Touch

Overview

Objectives

Implementation of spiking tactile system on a 7-DoF manipulator and development of manipulation algorithms for stable grasping of objects of different shape and softness. Tactile data will be integrated in the control loop of the manipulator to obtain optimal grasps taking the object shape, softness and deformability into account. Different types of end-effectors will be considered, i.e. from parallel grippers to under-actuated multi-fingered hands.

Expected Results

A new set of force controllers, i.e. impedance and admittance controllers, will be obtained allowing different robots, from mobile manipulators to humanoid robots, performing grasping of a large variety of objects. The performance of the new grasping algorithms will be compared to the performance obtained using grasping algorithms based on open-loop approaches and those relying on data provided by a wrist-mounted force/torque sensor. The goal is to show that the new algorithms based on the tactile system provide a much higher ratio of successful grasps considering a set of objects and different configurations. The demonstrator will feature a service robot being able to grasp a variety of objects involved in typical home daily tasks of elderly care scenarios

Secondments

  • IIT-iCub

    implementation of spiking tactile skin for robots

  • USFD

    learn predictive methods

  • UNIBI

    learn manipulation of switches

Supervisors

  • S. Terreri

  • C. Bartolozzi

  • L. Natale

  • T. Prescott

  • R. Haschke

Luca Lach

My interest in Robotics and Cognitive Science was sparked when I joined the RoboCup @Home Team TOBI. In 2017, we won the WorldChampionship in service robotics and for my Bachelor's Thesis I've work with one of these robots. During that time, I learned how robots drive, manipulate objects, form decision processes and how they can learn themselves. For my Master's Thesis, I explored the field of Reinforcement Learning in conjunction with Representation Learning. It was very exciting for me to acquire new skills in this field of research. My plan is to combine both of my main interest for my PhD.

Neutouch for me

What I like most about NeuTouch is the mix of many different sciences which try to achieve a common goal. The possibility to experience a different culture for three years motivated my choice as well. I hope that this time will create many interesting opportunities for exciting collaborations.

Info

  • Research Topics

    Robotics, machine learning

  • Institution

    PAL robotics

  • Background

    B.Sc. Cognitive Computer Science (2013-2017)
    M.Sc. Intelligent Systems (2017-2019)