Learning Object Recognition in a NeuroBotic System

Abstract

Object localisation and identification is a crucial problem for advanced mobile service robots. We developed an object recognition system that localises and identifies objects using a colour-based visual attention control algorithm and a hierarchical neural network for object classification utilising hierarchical class grouping. The approach is evaluated in a test scenario where a robot is situated in front of a table. The robot has to identify and manipulate objects lying on this table. We evaluated the total object recognition performance and compared the effectiveness of different feature sets. The approach showed very encouraging results and meets real-time constraints. 1