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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Detecting and localizing contacts is essential for robot manipulators toperform contact-rich tasks in unstructured environments. While robot skins canlocalize contacts on the surface of robot arms, these sensors are not yetrobust or easily accessible. As such, prior works have explored usingproprioceptive observations, such as joint velocities and torques, to performcontact localization. Many past approaches assume the robot is static duringcontact incident, a single contact is made at a time, or having access toaccurate dynamics models and joint torque sensing. In this work, we relax theseassumptions and propose using Domain Randomization to train a neural network tolocalize contacts of robot arms in motion without joint torque observations.Our method uses a novel cylindrical projection encoding of the robot armsurface, which allows the network to use convolution layers to process inputfeatures and transposed convolution layers to predict contacts. The trainednetwork achieves a contact detection accuracy of 91.5% and a mean contactlocalization error of 3.0cm. We further demonstrate an application of thecontact localization model in an obstacle mapping task, evaluated in bothsimulation and the real world.

Contact Localization for Robot Arms in Motion without Torque Sensing