Power efficient, bandwidth optimized and fault tolerant sensor management for IOT in Smart Home

This paper proposes a localized data processing and decision making framework for IoT enabled environment. It constitutes a master unit which efficiently manages local sensor networks in Smart Homes. This master unit collects data from the network of sensors installed in different locations within and around the house and intelligently identifies the dependencies among them. Based on this, the set of sensors are classified as dependent whose output values can be predicted using data from remaining of them. This knowledge extracted locally is shared with the Cloud Server, thereby providing a refined picture of the sensory data. On the basis of locally extracted knowledge, the sensors are tuned in real time to minimize the redundancy in usage and power consumption. Moreover, in cases of failure of a sensor unit, approximate decisions can still be made based on previous patterns.