A Sensor Networking Middleware for Clustering Similar Things
Tomohiro Nagata, Hisashi Oguma, Kenichi Yamazaki · 2015
Abstract. To realize context awareness, we have to retrieve context from the real world. The state of things around the user is a context-rich source of information. promising context. The goal of this research is to retrieve such information using a sensor network system in which micro sensors are attached to everything. We focus on the similarity between sensor values and propose a middleware module that creates sensor clusters based on similarity. The middleware separates universal mechanisms for cluster management and domain knowledge of similarity. We implement a prototype using MOTE, a three-axis accelerometor, and premilinary code to show its feasibility. 1