基于SOFM神经网络的无线传感器网络数据融合算法

Data Aggregation in WSN Based on SOFM Neural Network

为了降低无线传感器网络的通信量,降低能耗,延长网络的生命周期,提出了一种基于SOFM(Self-Organizing Feature Mapping)神经网络的数据融合算法(SOFMDA),该算法将自组织映射神经网络和无线传感器网络分簇路由协议相结合,使簇中的各个节点完成神经元的工作,按照数据的特征对其进行分类,提取同类数据的特征,将特征数据发送到汇聚节点,从而减少了数据发送量,延长网络的生命期。仿真实验表明,与普通的数据融合方法相比,SOFMDA能够在保证数据准确性的前提下,有效减少网络通信量,延长网络生命期。在文中仿真实验的时间内,达到了LEACH算法性能的1.5倍。

In order to reduce the communication traffic in wireless sensor networks and reduce energy consumption and increase the network lifetime, a data fusion algorithm based on SOFM （Self-Organizing Feature Mapping）neural network（ SOFMDA ）was proposed, which combined self-organizing neural networks and wireless sensor network clustering routing protocol. Each node in WSN performed is a neuron. According to the data characteristics, SOFMDA made classification, in which the data with same characteristic was classified into the same class, then got the feature data which is sent to the Sink node. Thereby reducing the amount of data sent, to extend the lifetime of the network. Simulation results show that compared with conventional data fusion methods, SOFMDA can guarantee the accuracy of the data under the premise of effectively reducing network traffic, extending the network lifetime. During the simulation,the performance of SOFMDA reached 150% of LEACH＇s.