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

Learning from demonstrations is a promising paradigm for transferringknowledge to robots. However, learning mobile manipulation tasks directly froma human teacher is a complex problem as it requires learning models of both theoverall task goal and of the underlying actions. Additionally, learning from asmall number of demonstrations often introduces ambiguity with respect to theintention of the teacher, making it challenging to commit to one model forgeneralizing the task to new settings. In this paper, we present an approach tolearning flexible mobile manipulation action models and task goalrepresentations from teacher demonstrations. Our action models enable the robotto consider different likely outcomes of each action and to generate feasibletrajectories for achieving them. Accordingly, we leverage a probabilisticframework based on Monte Carlo tree search to compute sequences of feasibleactions imitating the teacher intention in new settings without requiring theteacher to specify an explicit goal state. We demonstrate the effectiveness ofour approach in complex tasks carried out in real-world settings.

Combined Task and Action Learning from Human Demonstrations for Mobile Manipulation Applications