Adaptive cluster-based relay-node placement for disjoint wireless sensor networks

Abstract

Wireless sensor networks are formed with very small sensor devices with limited energy and short transmission range. Sensors are randomly deployed in remote areas with harsh conditions. Due to this, their distribution are non-uniform. In a hostile environment, the occurring of large scale failure is very often and it creates partitions of different sizes in the network. Restoring the network connectivity is very crucial for network functioning. Typically sensors are static and communication is bi-directional and we need to reconnect the network with the additional relay sensor nodes. After reconnecting the network, overhead on relays is very high and thus it consumes more energy and high computational power. Optimal number of relay nodes required for reconnecting the network is an NP-hard problem. Here, we proposed a new heuristic approach called adaptive Cluster-based Relay Placement Algorithm. Our approach is to reconnecting the networks based on finding the nearest partitions using their centroids. Using the proposed algorithm, relays are placed on the denser part of the partitions and number of required relays are also minimum. It reduces the network recovery time and increases the network lifetime. Through simulation, we show that the proposed algorithm performs better than the existing approaches.