Study On Self-health Status Fusion Mechanism For Wireless Sensor Network

The network constituted by a large number of sensor nodes which deployed randomly in the monitored area by means of wireless self-organization can be used for various of monitoring tasks cooperatively, such as battle field information collection, toxic gas monitoring, coal mine safety monitoring, and marine ecological monitoring. These environments are very complex, which may lead to instable work of the sensor nodes or even failure, so that can't competent for the monitoring task. By monitoring the self health information in wireless sensor network, the failure area can be known in time and corresponding actions can be taken, which assures the network recover from failure quickly to provide the monitoring task normally. However, wireless sensor networks are resource-constrained. If each node sends the monitored health information to the sink without any processing, the nodes, especially the nodes near the sink will forward a large number of packets, which will lead premature death of the node due to a large amount of energy consumption. So the aggregating of the self health status becomes necessary. This study comes from National Nature Science Foundation and focus on self health status aggregation.The thesis introduces the current state of the art of data fusion in wireless sensor network. For the characteristic of the self health information which is reported periodically by all the nodes, the thesis constructs an aggregation tree based on grid. The aggregation tree divides the network into grids according to the position information of the node, each grid forms a cluster. The cluster head is elected not only considers the residual energy, but also considers the position information. The cluster head will consume more energy due to take responsibility of data fusion and communicating with cluster members and other clusters. In order to avoid premature death of the cluster head and unnecessary energy consumption in reconstructing aggregation tree, the cluster head is replaced locally. Namely, the new cluster head will be reelected if the residual energy of the cluster head is less than half of the energy when it becomes cluster head.In the grid cluster, the thesis designs an aggregation algorithm that aggregates the coverage in the cluster and the link quality of the member nodes to cluster head based on the assumption that the information in a multi-node covered area can be detected and reported only if at least one of the nodes detects and reports it successfully,Simulation experiments indicate that the network consumes lesser energy and enjoys longer network lifetime through the aggregation tree. The self health status aggregation algorithm proposed in the thesis can accurately reflect the health information of the nodes in the cluster area.