TagFree: Identifying users without tags in smart home environments

Abstract

Since family members have their unique features when living in a smart home environment, user identifications are able to achieve without any tags. In this paper, we propose TagFree system in which users freely move in a smart home environment and TagFree system is able to intelligently identify family member according to sensed data. Specifically, TagFree system consists of two phases: the training phase and the prediction phase. In the training phase, sensed data are collected and then, given a huge amount of sensed data, the profile of users, including the most common sensed data (i.e., tones, weights and location), are discovered. Once the profile of users is built up, in the prediction phase, we propose two scoring algorithms to generate likelihood scores according to the sensed data given. A simulation is implemented to verify the correctness of our proposed system and extensive experiments are conducted. Experimental results show that our proposed TagFree is able to achieve high accuracy of identifying family member without any tags. Furthermore, from experimental results, we also provided some guidelines to set some important parameters for TagFree.