HiCHO: Attributes based classification of ubiquitous devices

Abstract

An online and incremental clustering method to classify heterogeneous devices in dynamic ubiquitous computing environment is presented. The proposed classification technique, HiCHO, is based on attributes characterizing devices. These can be logical and physical attributes. Such classification allows to derive class level similarity or dissimilarity between devices and further use it to extract semantic information about relationship among devices. The HiCHO technique is protocol neutral and can be integrated with any device discovery protocol. Detailed simulation analysis and real-world data validates the efficacy of the HiCHO technique and its algorithms. © 2012 ICST Institute for Computer Science, Social Informatics and Telecommunications Engineering.