Many of today’s robot perception systems aim at accomplishing perception tasks that are too simplistic and too hard . We propose a perception system that maintains its beliefs about its environment as a scene graph with physics simulation and visual rendering . We investigate avariation of robot perception tasks suitable for robots accomplishing everydaymanipulation tasks, such as household robots or a robot in a retail store . The result is a system that can provide useful information for manipulation tasks . Insuch settings it is reasonable to assume that robots know most objects and havedetailed models of them . Insuinging objects, the perception system retrieves the model of the object and places it at the corresponding place in a VR-based environment model. The physics simulation ensures that object detections that are physically not physically notpossible are rejected are rejected and scenes can be rendered to generate expectations at the image level. The result of the system can be used to create a perception systems that can be easily manipulated by robots. Insuing a robot can provide helpful information about the objects that are

Author(s) : Patrick Mania, Franklin Kenghagho Kenfack, Michael Neumann, Michael Beetz

Links : PDF - Abstract

Code :
Coursera

Keywords : perception - robot - tasks - system - robots -

Leave a Reply

Your email address will not be published. Required fields are marked *