Estimating which Object Type a Sensor Node is Attached to in Ubiquitous Sensor Environment

Background And Goal

Obtaining states and state changes of objects is becoming possible by attaching cheap sensor nodes to indoor objects (Intille et al., 2003). This has triggered surveillance studies concerned with end-user sensor installation. The following comments on end-user installation can be found in Beckmann (2004): ‘‘the monetary and time cost of professional installation is prohibitive for non-critical applications’’ and ‘‘leveraging the fact that an end-user is a domain expert for his own home can lead to an application better tailored to his needs or preferences.’’ However, some problems have arisen as regards end-user installation. For example, Beckmann et al. (2004) revealed that some end-users could not understand the meaning of the association itself in the experiment, where a bar-code reader and bar-codes attached to sensor nodes were used to associate a sensor node and the type of object to which the node is attached. Also, it takes an average of 84 minutes for participants to deploy ten sensor nodes in an experiment (this is not only the time needed for the association). If there are dozens of nodes, manual association will become a burden for end-users. Tapia et al. (2004) also mention that installation without association would dramatically reduce the installation time when engineers install the sensors in a home.