English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Several robot manipulation tasks are extremely sensitive to variations of thephysical properties of the manipulated objects. One such task is manipulatingobjects by using gravity or arm accelerations, increasing the importance ofmass, center of mass, and friction information. We present SwingBot, a robotthat is able to learn the physical features of a held object through tactileexploration. Two exploration actions (tilting and shaking) provide the tactileinformation used to create a physical feature embedding space. With thisembedding, SwingBot is able to predict the swing angle achieved by a robotperforming dynamic swing-up manipulations on a previously unseen object. Usingthese predictions, it is able to search for the optimal control parameters fora desired swing-up angle. We show that with the learned physical features ourend-to-end self-supervised learning pipeline is able to substantially improvethe accuracy of swinging up unseen objects. We also show that objects withsimilar dynamics are closer to each other on the embedding space and that theembedding can be disentangled into values of specific physical properties.

SwingBot: Learning Physical Features from In-hand Tactile Exploration for Dynamic Swing-up Manipulation