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

Traditional AI-planning methods for task planning in robotics require asymbolically encoded domain description. While powerful in well-definedscenarios, as well as human-interpretable, setting this up requires substantialeffort. Different from this, most everyday planning tasks are solved by humansintuitively, using mental imagery of the different planning steps. Here wesuggest that the same approach can be used for robots, too, in cases whichrequire only limited execution accuracy. In the current study, we propose anovel sub-symbolic method called Simulated Mental Imagery for Planning (SiMIP),which consists of perception, simulated action, success-checking andre-planning performed on 'imagined' images. We show that it is possible toimplement mental imagery-based planning in an algorithmically sound way bycombining regular convolutional neural networks and generative adversarialnetworks. With this method, the robot acquires the capability to use theinitially existing scene to generate action plans without symbolic domaindescriptions, while at the same time plans remain human-interpretable,different from deep reinforcement learning, which is an alternativesub-symbolic approach. We create a dataset from real scenes for a packingproblem of having to correctly place different objects into different targetslots. This way efficiency and success rate of this algorithm could bequantified.

Simulated Mental Imagery for Robotic Task Planning