Human activity recognition based on surrounding things

This paper proposes human activity recognition based on the actual semantics of the human's current location. Since predefining the semantics of location is inadequate to identify human activities, we process information about things to automatically identify the semantics based on the concept of affordance. Ontology is used to deal with the various possible representations of things detected by RFIDs, and a multi-class Naive Bayesian approach is used to detect multiple actual semantics from the terms representing things. Our approach is suitable for automatically detecting possible activities under a variety of characteristics of things including polysemy and variability. Preliminary experiments on manually collected datasets of things demonstrated its noise tolerance and ability to rapidly detect multiple actual semantics from existing things.