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

Multi-fingered robotic hands have potential to enable robots to performsophisticated manipulation tasks. However, teaching a robot to grasp objectswith an anthropomorphic hand is an arduous problem due to the highdimensionality of state and action spaces. Deep Reinforcement Learning (DRL)offers techniques to design control policies for this kind of problems withoutexplicit environment or hand modeling. However, state-of-the-art model-freealgorithms have proven inefficient for learning such policies. The main problemis that the exploration of the environment is unfeasible for suchhigh-dimensional problems, thus hampering the initial phases of policyoptimization. One possibility to address this is to rely on off-line taskdemonstrations, but, oftentimes, this is too demanding in terms of time andcomputational resources. To address these problems, we propose the A Grasp Poseis All You Need (G-PAYN) method for the anthropomorphic hand of the iCubhumanoid. We develop an approach to automatically collect task demonstrationsto initialize the training of the policy. The proposed grasping pipeline startsfrom a grasp pose generated by an external algorithm, used to initiate themovement. Then a control policy (previously trained with the proposed G-PAYN)is used to reach and grab the object. We deployed the iCub into the MuJoCosimulator and use it to test our approach with objects from the YCB-Videodataset. Results show that G-PAYN outperforms current DRL techniques in theconsidered setting in terms of success rate and execution time with respect tothe baselines. The code to reproduce the experiments is released together withthe paper with an open source license.

A Grasp Pose is All You Need: Learning Multi-fingered Grasping with Deep Reinforcement Learning from Vision and Touch