| 摘要: |
| 在研究现有定位算法的基础上，针对基于接收信号强度指示(RSSI)定位模型中的参数易受环境影响等问题，提出了一种新型的粒子群优化(PSO)算法与后向传播（BP）神经网络相结合的算法。BP网络算法权值的修正依赖于非线性梯度值，易形成局部极值，同时学习次数较多，需先通过粒子群算法进行优化。为了提高定位精度，首先采用速度常量法滤波处理，然后通过改进的混合优化算法对BP神经网络初始权值和阈值进行优化，并分析算法的性能。试验中隐层节点个数采用试错法，从12到19变化，以确定合适数目。实验结果表明，与一般加权算法和传统BP算法相比，改进的混合优化算法可大幅改善测距误差对定位误差的影响，同时可使25 m内最小定位误差小于0.27m |
| 关键词: 无线传感器网络 定位算法 测量误差 BP神经网络 粒子群优化 路径损耗模型 |
| DOI： |
|
| 基金项目:贵州省科技支撑计划(黔科合GZ字\[2015\]3034);国家自然科学基金资助项目(51475097)；国家科技支撑计划(2014BAH05F01)；贵州省科技基金项目(黔科合J字\[2015\]2043);贵州省基础研究重大专项(黔科合JZ字\[2014\]2001) |
|
| Optimization of wireless sensor network positioning based on PSO-BP algorithm |
| BIAN Guolong,HUANG Haisong,WANG Anyi,YU Kaihua |
| () |
| Abstract: |
| The existing localization algorithms are discussed.For the problem that received signal strength indication(RSSI) based on the parameter model of positioning is easily affected by environment,a novel algorithm is proposed which combines particle swarm optimization(PSO) algorithm with back propagation(BP) neural network. The correction of the weight of the BP network algorithm depends on the nonlinear gradient value. It is easy to form local extremum. At the same time,the number of learning is more. It should be optimized by PSO algorithm. In order to improve the positioning accuracy,the speed constant method is used to perform filtering. Then the initial weights and thresholds of the BP neural network is optimized by the improved hybrid optimization algorithm.The performance of the algorithm is compared with that of the existing positioning algorithms. The number of hidden layer nodes varies from 12 to 19. The experimental results show that the improved hybrid optimization algorithm can significantly improve the effect of ranging error on the positioning error compared with the general weighted algorithm and the traditional BP algorithm. The minimum positioning error can reach 0.27 m in 25 m range. |
| Key words: wireless sensor network(WSW) localization algorithm measurement error back propagation(BP) neural network particle swarm optimization(PSO) path loss model |
《电讯技术》编辑部 