Wireless Sensor Networks Unequal Clustering Pairs Of Cluster Head Selection Mechanism

In wireless sensor networks, we need to research a rational and efficient routing protocol algorithm, with a view to prolong the network lifetime and balance network energy consumption. At present, comparing with the plane routing, clustering routing algorithm is considered to have splendid development prospect, however, there are circs of excessive burden on cluster head node, the inequality energy consumption of the node,as well as forming the "hot spot" easy in most clustering routing algorithm. Several algorithms can alleviate the burden of the main cluster head effectively by introducing the idea of Double Cluster Heads, they have no method aiming at the selection mechanism of assisting cluster head and the issue of reducing the energy consumption which grows out of choosing cluster head by competition mechanism, thus the issues affected the functions of wireless sensor networks to some extent.In this thesis, we introduce double cluster heads mechanism which can share the burden of main cluster head by using assisting cluster head to take on the data switching, besides, for the purpose of reducing the frequency of clustering, prolonging the lifecycle of cluster and reducing the energy consumption which grows out of cluster head selection competition, we introduce the Algorithm of Choosing the Uneven Distributed Double Cluster Heads in WSN Based on ARMA Traffic Prediction(AUDC).The algorithm can predict the next moment traffic of assisting cluster head by using the combination of the main cluster head and ARMA Traffic Prediction, and then we choose the assisting cluster head according to the residual energy to secure the ends of replacing Vice-cluster head timely. In order to solve the network "hot spot" problem effectively, we bring forward Uneven Clustering Algorithm of WSN based on Cirque Subarea with Different Bandwidth (UCCB), the algorithm partitions the network area by using cirque which has unequal bandwidths and fan-shaped area, We can take the optimal value of the number of different partitions, as well as the growth factor of bandwidth and so on by using the boundary conditions, There should be a uneven clustering,all above-mentioned is in order to balance energy consumption of the network nodes and make the total energy consumption minimum.With the help of MATLAB platform, we simulate the two kinds of new algorithm and compare them with LEACH, HEED and so on, the result shows that the new algorithms can prolong the network lifecycle effectively and realize the load balancing in the network.