Object recognition based on depth information and associative memory

Author

Puls, Stephan ; Schnorr, N. ; Worn, Heinz

Author_Institution

Inst. of Process Control & Robot, Karlsruhe Inst. of Technol., Karlsruhe, Germany

fYear

2012

fDate

16-18 Nov. 2012

Firstpage

186

Lastpage

191

Abstract

Steady improvement of robotic systems due to developments in the realm of sensing the world enables advances towards human-robot-cooperation. In order for the robot to be reactive in its environment objects need to be identified. In this paper an approach is presented which allows identification of objects in the working area of an industrial robot. Neural Networks are used as associative memory to learn new items and efficiently recognize learned objects.

Keywords

Hopfield neural nets; content-addressable storage; human-robot interaction; industrial robots; learning (artificial intelligence); object recognition; production engineering computing; robot vision; Hopfield nets; associative memory; depth information; human-robot-cooperation; industrial robot; neural networks; object identification; object recognition; robotic systems; Biological neural networks; Humans; Object recognition; Robot sensing systems; Service robots; Object recognition; associative memory; cognitive robotics; depth information; neural network;

fLanguage

English

Publisher

ieee

Conference_Titel

Robotic and Sensors Environments (ROSE), 2012 IEEE International Symposium on

Conference_Location

Magdeburg

Print_ISBN

978-1-4673-2705-3

Type

conf

DOI

10.1109/ROSE.2012.6402606

Filename

6402606