基于MFOA-GRNN模型的三维定位研究  

THREE-DIMENSIONAL POSITIONING BASED ON MFOA-GRNN MODEL

作　　者：马翠红 徐天天 杨友良 Ma Cuihong;Xu Tiantian;Yang Youliang(College of Electrical Engineering,North China University of Science and Technology,Tangshan 063210,Hebei,China)

机构地区：[1]华北理工大学电气工程学院,河北唐山063210

出　　处：《计算机应用与软件》2020年第7期212-215,280,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61171058)。

摘　　要：由于传统室内定位模型容易受到外界因素影响导致定位精度大幅下降,提出一种将多种群果蝇优化算法(Multi-population Fruit Fly Optimization Algorithm,MFOA)与广义回归神经网络(Generalized Regression Neural Network,GRNN)相结合的MFOA-GRNN三维室内定位模型。基于射频识别技术并引入多种群的果蝇优化算法用以选择GRNN的平滑参数,并通过MFOA-GRNN模型将阅读器接收信号强度与目标坐标进行对应进而判断目标位置。仿真结果表明,该模型克服了传统室内定位模型受主观因素影响大、学习效率低等不足,同时使算法的全局寻优能力得到了加强,定位精度显著提高。Traditional indoor positioning model is vulnerable to external factors,which leads to the great decline in positioning accuracy.Therefore,this study proposes a three-dimensional indoor positioning model based on MFOA-GRNN,which combines multi-population fruit fly optimization algorithms(MFOA)and generalized regression neural network(GRNN).It was based on RPID and introduced multi-population fruit fly optimization algorithm to select the smooth parameters of GRNN.In addition,through MFOA-GRNN model,the signal strength received by the reader was corresponding to the target coordinate,so as to determine the target position.The simulation experiments show that this model overcomes the shortcomings of traditional indoor positioning model,which is influenced by subjective factors and low learning efficiency.The global optimization ability of the algorithm has been strengthened,and the positioning accuracy has been significantly improved.[HJ]

关 键 词：广义回归神经网络 室内定位 MFOA-GRNN 改进果蝇算法 射频识别技术 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]