Study On Clustering Routing Algorithm Based On PAM For Wireless Sensor Networks

In wireless sensor networks,nodes have limited energy and cannot be charged.In order to solve this problem,clustering routing algorithm is adopted to optimize the network energy consumption,collect data and improve the effective life of the network.Although clustering routing algorithm has certain advantages,there are still some important challenges in selecting sensor as cluster head,which has an important impact on improving energy efficiency.In the clustering stage,the nodes are divided into several clusters,and then a number of sensor nodes are selected as the head of each cluster.In a typical cluster-type wireless sensor network,the wireless sensor node monitors the data and sends the monitored data to the cluster-head node,which collects and aggregates the data and sends it to the base station.Node clustering in wireless sensor networks has the advantages of scalability,energy saving and reducing routing delay.In this paper,several clustering methods are studied and their advantages and disadvantages are illustrated.In the analysis and research of the machine semester algorithm,based on the partition clustering algorithm,the traditional clustering routing algorithm is improved.By introducing the intelligent algorithm,the selection of cluster head node is optimized to avoid the result falling into the local optimal situation.The main work is as follows:(1)In the routing algorithm of wireless sensor network based on k-medoids,the principle of SECA algorithm is analyzed.Through detailed performance analysis,it is found that the algorithm is easily affected by extreme values.Therefore,on this basis,we improve its shortcomings and use k-medoids algorithm as the core algorithm of wireless sensor network clustering.At the same time,we use the centralized method to optimize the selection of initial cluster-head nodes.By calculating the central circle and selecting the initial cluster-head nodes uniformly on the circle,it can obviously reduce the number of iterations,accelerate the network initialization time,and solve the problem that the extreme value is easy to affect the overall network survival time.(2)In the study of clustering routing algorithm for adaptive wireless sensor networks based on nearest-neighbor propagation,the concept of nearest-neighbor propagation is introduced to optimize the selection of initial cluster-head nodes.We combine the k-medoids algorithm with the nearest neighbor propagation algorithm to achieve better clustering.Our proposed algorithm consists of two steps.Step 1:using the nearest neighbor propagation algorithm,the initial cluster head is automatically selected according to the network stitching.In this step,we do not need to calculate the number of cluster heads in advance.According to the nearest neighbor propagation algorithm,the node adaptively selects the appropriate cluster head.Step 2:cluster the network.The k-medoids algorithm is used to obtain the final clustering results.In this step,we use the initial cluster-head node as the parameter for the initial iteration.It greatly reduces the iteration time and avoids the result falling into the local optimal solution.