In this work we propose an approach to learn a robust policy for solving the pivoting task. Recently, several model-free continuous control algorithms were shown to learn successful policies without prior knowledge of the dynamics of the task. However, obtaining successful policies required thousands to millions of training episodes, limiting the applicability of these approaches to real hardware. We developed a training procedure that allows us to use a simple custom simulator to learn policies robust to the mismatch of simulation vs robot. In our experiments, we demonstrate that the policy learned in the simulator is able to pivot the object to the desired target angle on the real robot. We also show generalization to an object with different inertia, shape, mass and friction properties than those used during training. This result is a step towards making model-free reinforcement learning available for solving robotics tasks via pre-training in simulators that offer only an imprecise match to the real-world dynamics. ; QC 20171023


    Access

    Download


    Export, share and cite



    Title :

    Reinforcement Learning for Pivoting Task


    Contributors:

    Type of media :

    Paper


    Type of material :

    Electronic Resource


    Language :

    English



    Classification :

    DDC:    629



    PIVOTING SYSTEM, PIVOTING ASSEMBLY, PIVOTING MODULE AND SEAT COMPRISING THE PIVOTING SYSTEM

    GAROTTE GÉRALD / TOUZET BERTRAND | European Patent Office | 2021

    Free access

    Pivoting system, pivoting assembly, pivoting module and seat comprising the pivoting system

    GAROTTE GÉRALD / TOUZET BERTRAND | European Patent Office | 2021

    Free access

    Pivoting door

    DIERINGER MARK S / BORNHORST DOUGLAS M / GILLILAND KEVIN A | European Patent Office | 2018

    Free access

    PIVOTING DOOR

    European Patent Office | 2022

    Free access

    PIVOTING FURNITURE

    MILLIS JOHN / GROSHONG JASON / BERNS MICHAEL | European Patent Office | 2018

    Free access