Boundary And Bottleneck Detection Based On Topology Information In Wireless Sensor Networks

Wireless Sensor Network(WSN)is composed of a large number of low-cost sensors which can collect and forward data.WSN has been widely used in the fields of environment,health monitoring and intrusion detection.Since the randomness of sensor deployment will affect the overall performance of the network,and the optimization of network requires the first acquisition of the overall information.The boundary and bottleneck are among the most important overall messages.It is found that the number of the shortest path between different node pairs rarely passes through the boundary nodes,that is,the betweenness centrality of the boundary is less than that of other nodes.Therefore,betweenness centrality is proposed to study the boundary and bottleneck of WSN for the first time in this paper.(1)In this paper,an algorithm is proposed to identify the boundary of a wireless sensor network composed of a large number of sensor nodes with good average degree.The algorithm is based on the betweenness centrality of nodes and only depends on the connection information of nodes.Due to the large number of sensor nodes in the real network,the accurate calculation of the betweenness centrality of each point consumes huge resources and is not of practical significance.Therefore,the traversal number of the shortest path is greatly reduced by constructing several hop neighborhood graphs of nodes to approximate the value of the betweenness centrality.(2)However,when comparing betweenness centrality of sensors still need to collect global information,energy consumption is too large for the shortage of the resources on the sensors.To this end,this paper proposes an improved distributed method which only rely on the local network information.It was found that boundary could be recognized by narrowing the comparison range within the 5-hop neighborhood graph when comparing the betweenness centrality of nodes.This method transforms the global issues into a local solution and can reduce the energy consumption of the sensors.The algorithm screens out the mistakes existing in the candidate boundary nodes by judging whether the adjacent nodes with boundary nodes in the end.The simulation results show that the algorithm has good applicability and can provide a precondition for bottleneck nodes identification.(3)A distributed algorithm is proposed in order to identify the bottleneck nodes.The method is based on boundary detection and only depends on node connectivity.Inspired by the idea of dilation and erosion,the detected boundary is used as the initial nodes to eroded the whole network continuously,and then the last corroded nodes are used as the initial nodes to expand outward continuously.The nodes at the junction of different regions are the bottleneck nodes.In this topic,the proposed algorithm for boundary and bottleneck nodes detection is easy to understand,and the application in different target deployment areas and different node densities is discussed through simulation.The results verify the feasibility of the algorithm and provide effective information for other optimization methods.