一种基于GWO-BP神经网络的RSSI测距算法  

A RSSI ranging algorithm based on GWO-BP neural network

作　　者：林贻若 余科根 朱飞洋 布金伟 LIN Yiruo;YU Kegen;ZHU Feiyang;BU Jinwei(School of Environmental Science and Spatial Informatics,China University of Mining and Technology,Xuzhou 221116,China;Faculty of Land Resources Engineering,Kunming University of Science and Technology,Kunming 650093,China)

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116 [2]昆明理工大学国土资源工程学院,云南昆明650093

出　　处：《测绘学报》2024年第8期1564-1573,共10页Acta Geodaetica et Cartographica Sinica

基　　金：中国矿业大学研究生创新计划(2024WLKXJ174);中央高校基本科研业务费专项(2024-10980);江苏省研究生科研与实践创新计划项目(KYCX24_2825);国家自然科学基金(42174022)。

摘　　要：近些年来,基于接收信号强度指示(RSSI)测距的研究受到了广泛的关注,特别是在物联网和室内定位领域。精准的距离估计是基于测距方法实现高精度定位的基础,但由于存在测量噪声和多路径效应的影响,RSSI信号具有很强的波动性,从而导致RSSI与空间真实物理距离之间的映射关系具有不均匀的特点。为了增强RSSI与真实物理距离之间的映射关系,提高RSSI测距的精度,本文基于反向传播(BP)神经网络和灰狼优化(GWO)算法,提出了一种基于GWO-BP神经网络的RSSI测距算法。GWO算法相比于粒子群优化算法(PSO)、遗传算法(GA)、差分演化(DE)、进化编程(EP)、进化策略(ES)具有更快收敛速度和更强稳定性的特点。此外,本文通过开发的手机软件采集实测数据,通过在两个不同的环境内进行试验。结果表明,基于路径损耗模型(PLM)测距的均方根误差(RMSE)分别为2.218、2.059 m,传统BP神经网络测距算法的RMSE分别为1.541、1.551 m,基于GA算法优化BP神经网络测距算法的RMSE分别为1.269、1.201 m,本文提出的GWO-BP神经网络测距算法的RMSE分别为1.054、0.833 m;结果表明本文算法测距精度更高,稳健性更好。Recently,the research on received signal strength indication(RSSI)based ranging has received a significant attention,especially in the field of Internet of things and indoor positioning.Precise distance measurement is the basis for high-precision positioning based on ranging algorithms,but the RSSI signal is highly fluctuating due to measurement noise and multi-path effects,which leads to a non-uniform mapping relationship between RSSI and the real physical distance in space.In order to enhance the mapping relationship between RSSI and real physical distance and improve the precision of RSSI ranging,this paper proposes a RSSI ranging algorithm based on GWO-BP neural network,which makes use of back propagation(BP)neural network and gray wolf optimization(GWO)algorithm.GWO algorithm has faster convergence and greater stability than particle swarm optimization(PSO),genetic algorithm(GA),differential evolution(DE),evolutionary programming(EP)and evolution strategy(ES).Furthermore,in this paper,the results of the experiments conducted in two different environments by collecting real data through the developed smartphone software show that:the root mean square error(RMSE)of the path loss model(PLM)based ranging were 2.218,2.059 m,the RMSE of the traditional BP neural network ranging algorithm were 1.541,1.551 m,and the RMSE of the GA algorithm-based optimized BP neural network ranging algorithm were 1.269,1.201 m,respectively,and the RMSE of the GWO-BP neural network ranging algorithm proposed in this paper were 1.054,0.833 m,respectively.The results indicate that the RSSI ranging algorithm proposed in this paper has higher ranging precision and better robustness.

关 键 词：路径损耗模型 优化算法 BP神经网络 RSSI测距 室内定位 

分 类 号：P228[天文地球—大地测量学与测量工程]