Handling visual features losses during a coordinated vision-based task with a dual-arm robotic system

Renliw Fleurmond (1) , Viviane Cadenat (1)

1 LAAS-RAP - Équipe Robotique, Action et Perception

We address the problem of total visual features losses during a coordinated vision-based task. We present a new method allowing to reconstruct these features even if the image is completely unavailable. Contrary to previous works, the proposed technique allows to deal with moving objects characterized by both points and lines features. Simulation results show the relevance of the proposed approach.

Domaines: Automatique / Robotique

Renliw Fleurmond, Viviane Cadenat. Handling visual features losses during a coordinated vision-based task with a dual-arm robotic system. European Control Conference, Jun 2016, Aalborg, Denmark. ⟨hal-01959491⟩