萤火虫算法优化神经网络的无线传感器网络数据融合

Firefly Algorithm Optimized Neural Network Data Fusion in Wireless Sensor Network

为了降低无线传感器网络中收集数据的冗余性,提高网络的生存周期,并且针对无线传感器网络数据融合算法中使用单层BP神经网络与SOFM神经网络收敛慢、易出现局部最优解的缺点,设计了一种萤火虫算法优化神经网络的无线传感器网络数据融合的策略FA-BPNN。首先,每个簇首节点接收该簇内感知节点监测到的数据,并根据相关性,提取相关的特征数据;然后,依据萤火虫算法优化BP神经网络进行数据融合;最后,通过仿真实验对其可行性进行测试。仿真结果表明,FA-BPNN算法提高了网络数据融合的效率,减少了网络的能量消耗,延长了网络的生命周期。

In order to reduce the redundancy of data collection in WSN and improve life-cycle of network,due to data fusion in WSN using single-layer BP neural network and SOFM neural network convergence slowly and is prone to disadvantages of local optima,a firefly algorithm optimizing neural network data fusion strategy in WSN called FA- BPNN was designed. First,each cluster head node receives monitoring data from nodes within cluster-aware,and extracts relevant characteristics data based on relevance. Then,data was fused according to the firefly algorithm optimization BP neural network. Finally,simulation experiments was used to test its feasibility. The simulation results show that FA-BPNN improves the efficiency of data fusion,reduces the energy consumption and prolongs the life-cycle in the WSN.