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

This paper presents c2g-HOF networks which learn to generate cost-to-gofunctions for manipulator motion planning. The c2g-HOF architecture consists ofa cost-to-go function over the configuration space represented as a neuralnetwork (c2g-network) as well as a Higher Order Function (HOF) network whichoutputs the weights of the c2g-network for a given input workspace. Bothnetworks are trained end-to-end in a supervised fashion using costs computedfrom traditional motion planners. Once trained, c2g-HOF can generate a smoothand continuous cost-to-go function directly from workspace sensor inputs(represented as a point cloud in 3D or an image in 2D). At inference time, theweights of the c2g-network are computed very efficiently and near-optimaltrajectories are generated by simply following the gradient of the cost-to-gofunction. We compare c2g-HOF with traditional planning algorithms for variousrobots and planning scenarios. The experimental results indicate that planningwith c2g-HOF is significantly faster than other motion planning algorithms,resulting in orders of magnitude improvement when including collision checking.Furthermore, despite being trained from sparsely sampled trajectories inconfiguration space, c2g-HOF generalizes to generate smoother, and often lowercost, trajectories. We demonstrate cost-to-go based planning on a 7 DoFmanipulator arm where motion planning in a complex workspace requires only 0.13seconds for the entire trajectory.

Cost-to-Go Function Generating Networks for High Dimensional Motion Planning