Understanding of Metric-Topological maps for mobile robot navigation

Abstract

The research is presenting a technique through which to learn, hence understand mobile robot Metric-Topological navigation maps for the purpose of much understanding of navigated environments. The adopted learning technique is based on using Principles Components Analysis (PCA) technique. PCA is used to reduce navigated maps dimensionality, capture maps only important details, hence to learn inherent details and characteristics of the environment. Navigation maps were created as based on using a stereo vision measurement techniques (VSLAM), Al-Mutib et al. [1]. Maps sizes are fixed, however their inside details are not static, as the environment is a moving dynamic space. The adopted technique was useful in terms of learning and understanding the environments inherent characterizations. This will help to support an enhanced and improved mobile navigation.