| 摘要: |
| 针对接收信号强度指示(Received Signal Strength Indication,RSSI)定位模型易受环境影响导致测距误差较大的问题，提出了采用天牛须搜索(Beetle Antennae Search,BAS)优化后向传播（Back Propagation，BP）神经网络拟合测距模型，克服了对数衰减模型易受环境干扰、参数取经验值等问题。首先，利用卡尔曼滤波对RSSI值进行校正，将校正后的数据输入BAS-BP网络拟合出测距模型并通过测距模型输出距离值；然后，利用极大似然估计法求解未知节点的坐标。实验结果表明，与BP模型和粒子群优化的BP模型相比，改进方法收敛速度快，定位精度提高更加明显。 |
| 关键词: 无线传感器网络 RSSI测距定位 天牛须搜索 BP神经网络 极大似然估计 卡尔曼滤波 |
| DOI： |
|
| 基金项目:河南省高等学校重点科研项目（15A520109）；河南省科技攻关项目（172102210064） |
|
| An improved RSSI ranging location algorithm |
| ZOU Dongyao,CHEN Pengwei,LIU Kuan |
| (School of Computer and Communication Engineering，Zhengzhou University of Light Industry，Zhengzhou 450000，China) |
| Abstract: |
| The received signal strength indication(RSSI) model is vulnerable to environmental impact,which results in a large ranging error.For this problem,a model combining back propagation(BP) neural network with beetle antennae search(BAS) is proposed.This model overcomes the environmental interference and the problem of taking empirical value as parameter,which exist in the logarithmic attenuation model.The RSSI value is corrected by Kalman filtering.Then,the distance measurement model,which can output the distance value,is fitted by BAS-BP with the corrected data.The coordinates of unknown node is solved by the maximum likelihood estimation method.The improved method is compared with BP model and particle swarm optimization-BP(PSO-BP) model,respectively.The experimental results show that the improved method has a faster convergence rate and a more accurate positioning accuracy. |
| Key words: wireless sensor network RSSI ranging location beetle antennae search(BAS) BP neural network maximum likelihood estimation Kalman filtering |
《电讯技术》编辑部 