Based On Neural Network And Ant Colony Algorithm Of Wsn Clustering Routing Algorithm Research

The network bandwidth, sensor nodes’ energy and computing capability are limited in wireless sensor network. Therefore it’s very important to design routing algorithms that can save node energy consumption and extend the network lifetime. Routing algorithms can be divided into flat routing algorithms and clustering routing algorithms based on the structure formed by sensor nodes. Clustering routing algorithms are better than flat algorithms in management and scalability. They are suitable for large-scale and distributed wireless sensor network, so a lot of researchments are made on them. This paper proposes two improved algorithms combined with neural network and ant colony algorithm to solve the problems in cluster-head choice and data transmission phase. The paper’s research contents are as follows:Firstly, a clustering routing algorithm based on neural network is proposed. In the phase of cluster-head choose, a neural network model is established in the base station. The input vectors are the nodes which become cluster-heads through the adaptive learning of neural network. Each input vector consists of three elements:the node residual energy, the distance between nodes and base station, the number of neighboring nodes. A formula is established in the data transfer phase according to the residual energy and distance of neighboring cluster-heads. Cluster-heads can select the most appropriate next hop according to the formula. When the node’s density is small, cluster-heads can choose intermediate nodes for data transmission. The simulation results show that compared with related algorithms the new algorithm can save and balance energy consumption and prolong network lifetime.Secondly, a clustering routing algorithm based on ant colony algorithm is proposed. In the data transfer phase, an improved ant colony algorithm is integrated into the routing algorithm in order to find the best path. In the algorithm the pheromone and evaporation coefficient are optimized. Cluster-heads not only consider the remaining energy of the neighboring nodes but also consider the distance and data transmission direction between them when they select the next hop. The new algorithm improves the convergence speed and the global search capability of the traditional ant colony algorithm, reduces the energy consumption.The two improved algorithms all belong to clustering routing algorithm, but they have their own advantages. The first algorithm is suitable for cluster-head choice phase while the second algorithm is suitable for data transmission phase. A large number of simulation experiments are made in the same environment. The simulation results show that the two algorithms are able to save and balance energy consumption and prolong the network lifetime.