Estimated replication of data in wireless sensor networks

A wireless sensor network (WSN) can be envisaged as a distributed system where the sensor nodes (clients) continuously update the sensed information to the sink (server). In many real-time environments exact replica of data is not required by the end application to deduce the result. Data estimation/prediction techniques at the sink can be used to infer the desired output. The proposed design is an improvement of the model as proposed by Santini which uses the traditional LMS algorithm to predict the desired output at the sink. In this work we solve the synchronization of mode problem in the Santinipsilas model by effectively utilizing the the principles of adaptive filter theory. The focus of this paper is to make better use of communication resources in data replication platforms than approaches based on periodic refreshing.