Catching objects in-flight is an outstanding challenge in robotics. In this paper, we present a closed-loop control system fusing data from two sensor modalities: an RGB-D camera and a radar. To develop and test our method, we start with an easy to identify object: a stuffed Piglet. We implement and compare two approaches to detect and track the object, and to predict the interception point. A baseline model uses colour filtering for locating the thrown object in the environment, while the interception point is predicted using a least squares regression over the physical ballistic trajectory equations. A deep learning based method uses artificial neural networks for both object detection and interception point prediction. We show that we are able to successfully catch Piglet in 80% of the cases with our deep learning approach.


    Access

    Download


    Export, share and cite



    Title :

    Learning to Catch Piglets in Flight


    Contributors:

    Publication date :

    2020


    Remarks:

    Fast Neural Perception and Learning for Intelligent Vehicles and Robotics workshop at IROS 2019


    Type of media :

    Preprint


    Type of material :

    Electronic Resource


    Language :

    English




    Catch joint prevents flight of pig

    Haas, W.B. / Cunningham, L.J. | Engineering Index Backfile | 1952


    HANDLE CATCH

    European Patent Office | 2023

    Free access

    ADJUSTABLE CATCH

    ZUBAREV VLADIMIR IVANOVICH / GORDEJ SERGEJ EVGENEVICH / EVSEEV DMITRIJ YUREVICH | European Patent Office | 2020

    Free access

    HANDLE CATCH

    SEGAWA SHIRO | European Patent Office | 2018

    Free access

    CARGO CATCH

    FERRIGGI JR RICHARD H | European Patent Office | 2015

    Free access