Robust Sensor Network Deployment with Priority Based on Failure Centrality

In a wireless sensor network, low-cost sensor nodes monitor nearby physical phenomenon, process the data and communicate with each other wirelessly. This type of network is subject to failures, one of which is caused by dynamic changes in network topologies. To be more resilient to failures, network topologies should have alternative paths for routing protocols to deliver data to sinks. We use the concept of centrality to improve the robustness of sensor network deployment. We introduce failure centrality, a novel centrality index to measure the importance of a sensor node to network connectivity and packet delivery efficiency. A sensor node is important to network deployment if it has high centrality. This node, if fails, can disconnect the network or cause late delivery of packets. By using the centrality index, we can find the most critical nodes and provide backups. We can also trade off deployment cost for network robustness by providing more or less backups depending on our budget.