An ambient intelligence architecture for extracting knowledge from distributed sensors
Alessandra De Paola, Salvatore Gaglio, Giuseppe Lo Re, Marco Ortolani · 2009
Precisely monitoring the environmental conditions is an essential requirement for AmI projects, but the wealth of data generated by the sensing equipment may easily overwhelm the modules devoted to higher-level reasoning, clogging them with irrelevant details. The present work proposes a new approach to knowledge extraction from raw data that addresses this issue at different levels of abstraction. Wireless sensor networks are used as the pervasive sensory tool, and their computational capabilities are exploited to remotely perform preliminary data processing. A central intelligent unit subsequently extracts higher-level concepts represented in a geometrical space and carries on symbolic reasoning based on them. The same tiered architecture is replicated in order to provide further levels of abstraction.