A study on boundary detection in wireless sensor networks

In wireless sensor networks (WSN), a group of sensor nodes are distributed to oversee an area round the clock so that in case of an unnatural event, like forest fire, earth quake, air pollution, oil spill, etc., the event can be reported and the affected area can be located and estimated immediately. In self-organized large WSN with poor wireless connectivity, it is an important issue to optimize the amount of traffic generated in the network, to achieve the required accuracy of area estimation in real time keeping the communication and hence the energy requirement limited. In this article, we present an in-depth study of the existing models, and the cutting-edge techniques evolved so far based on these models, to solve the problem. It exposes the limitations of the existing sensing models and demands further research to develop appropriate realistic models for sensing. A thorough comparative study of various approaches enables us to find the most befitting one to achieve high precision of area estimation, for a specific application with predefined conditions of node deployment and node characteristics.