Relocating Redundant Sensors in Randomly Deployed Wireless Sensor Networks
Sakil Ahmed Chowdhury, Abderrahim Benslimane · 2018
In a Wireless Sensor Network (WSN), issues like redundant nodes and duplicated data resulting from deploying nodes within each other's sensing area are still a challenge for random deployment strategy. This paper investigates the issue of identifying and relocating redundant nodes in a WSN where sensors are tracking events. First, a directed bipartite network, which consists of two types of vertices - nodes and detected events, and edges joining nodes with events, is constructed from raw sensory data to identify redundant nodes. Next, these redundant nodes are either relocated or put into sleep mode using circle packing technique to enhance coverage while minimizing energy usage during relocation. Simulation results show that the proposed algorithm can identify approximately 10% redundant nodes while detecting 98% events accurately. Our proposed redundant node identification algorithm has runtime complexity of O(U) + O(E), where U is the number of sensors and E is the total number of edges, the algorithm is also able to increase efficiency in back-end data processing by reducing approximately 20% duplicate data. Results also showed that, our redundant node relocation strategy improves coverage and energy efficiency of the network.