基于烟花优化的WSN数据融合算法

Data fusion of WSN based on fireworks algorithm optimization

为减少无线传感器网络(WSN)中的数据冗余,延长网络生命周期,克服基于BP(反向传播)神经网络的数据融合算法收敛慢、数据精度低和易陷入局部最优的不足,提出一种基于烟花算法优化的无线传感器网络融合算法(IFWABP).首先在烟花算法中使用Tent混沌图改进烟花种群的初始位置分布,使初始烟花分布更加均匀;然后利用改进的烟花算法优化BP神经网络的权重矩阵和阈值矩阵等参数进行数据融合;最后通过仿真对算法的性能进行测试和分析.仿真结果表明:对比其他算法,IFWABP算法提升了WSN数据融合的精度,降低了网络能耗,延长了网络生命周期.

To reduce the redundancy of data in wireless sensor network(WSN),prolong the network life cycle and overcome the disadvantages of slow convergence,low parameter accuracy and easy falling into local optimal values of data fusion algorithm based on BP(back propagation)neural network,a data fusion algorithm for wireless sensor network optimized by fireworks algorithm(IFWABP)was proposed.First,Tent chaotic map was used to improve the initial position distribution of fireworks population in fireworks algorithm,and make the initial fireworks distribution more uniform.Then,the improved fireworks algorithm was used to optimize the parameters of BP neural network such as weight matrix and threshold matrix for data fusion.Finally,the performance of the algorithm was tested and analyzed through simulation experiments.Simulation results show that compared with other algorithms,IFWABP algorithm improves the accuracy of WSN data fusion,reduces the network energy consumption,and prolongs the network life cycle.