Certain investigations to lifetime enhancement algorithms for event detection in wireless sensor networks

Abstract

Wireless Sensor Networks (WSN) are finding wide range of applications in day to day life. The wireless sensor nodes present in the network are often limited in functionality due to its limitations in its energy source, computational capacity and transmission range. The lifetime of the nodes are determined by its energy source while improvisation of lifetime can be done by employing multiple energy efficient techniques. Most of the sensor nodes are powered by non-rechargeable batteries. Energy efficient protocols that consider self organization, optimization and distributed co-ordination enhance the lifetime of the sensor nodes. Rare event detection is an important application of WSN in which there is a tradeoff between the network longevity and the rare event detection delay. For the research, the rare events that are long lived are only considered with the rare event detection probability closer to one. The research is pursued to detect rare events and get it communicated to the base station or the sink node within the guaranteed detection delay. The research is aimed at designing algorithms for lifetime enhancement of WSN detecting rare events. The main consideration is to reduce the energy consumption of radio transceivers, by employing multiple lifetime enhancement techniques like data aggregation, in-network processing and duty cycling of sensor nodes. For the first work of research, two dimensional sensor field is chosen. An heuristic approach based on 4*4 cluster size is proposed to reduce the hotspot problem based on Data Aggregation Through Alternate True (DATAT) structure. Results indicate reduced hotspot by employing alternate tree structures for data collection in alternate cycles. The simulation results show that DATAT improves the lifetime of the nodes by 5.5% over single hop communication and 5.9% over single tree based data aggregation. The DATAT is also found to have inbuilt fault tolerance behaviour but its usage is limited to 4*4 cluster newline