基于BP神经网络和蚁群的WSN分簇算法的研究

摘要

在研究经典低能量自适应分簇路由算法的基础上,提出基于BP神经网络和蚁群的无线传感器网络分簇路由算法.在分簇阶段将BP神经网络应用于每个分簇结构中,把采集的大量信息通过设计好的神经网络模型融合得到反映原始数据特征的少量数据信息.在数据传递阶段将蚁群算法应用到簇间路由机制中,寻找簇头到基站的最佳路径.将融合后的数据通过优化的路径传送到汇聚节点.仿真结果表明,该算法和LEACH算法相比,有效地均衡了网络的能量消耗,并延长了网络的生命周期.%On the basis of studying the classic low energy adaptive clustering hierarchy(LEACH)algorithm,the clustering routing algorithm of wireless sensor network(WSN)based on BP neural network and ant colony algorithm(ACA)is proposed. In clustering stage,BP neural network is applied to each clustering structure,and a large amount of acquired information is tran-sited to the designed BP neural network model. A little data information reflecting original data features is obtained by fusion. In data transmission stage,ACA is applied to inter-cluster routing mechanism to search the optimal path from the cluster head to the base station. The fused data is transited to clustering node through the optimized path. The simulation results indicate that in comparison with LEACH algorithm,the proposed algorithm can balance network energy consumption effectively,and lengthen the network life cycle.