一种基于抗差自适应的交互式多模型定位算法  

A interacting multi-model localization algorithm based on robust adaptive

作　　者：田步高 程龙[1,2] 刘杰民 TIAN Bugao;CHENG Long;LIU Jiemin(Key Laboratory Inelligent Information Processing&Microwave Technology,Northeasterm University at Qinghuangdao,Qinhuangdao,Hebei 066004,China;School of Computer and Communication Engineering,Notheasterm University at Qinhuangdao,Qinbuangdao,Hebei 066004,China)

机构地区：[1]东北大学秦皇岛分校智能信息处理与微波技术重点实验室,河北秦皇岛066004 [2]东北大学秦皇岛分校计算机与通信工程学院,河北秦皇岛066004

出　　处：《燕山大学学报》2023年第3期200-207,215,共9页Journal of Yanshan University

基　　金：河北省自然科学基金资助项目(F2020501012)。

摘　　要：针对无线传感器网络研究中距离信号传播过程存在的非视距现象,假定无线网络中的视距和非视距信道切换是一种具有两种状态的马尔可夫过程,提出了一种基于抗差自适应的交互式多模型算法,基于蜂窝网络到达时间和信号衰减值,利用抗差自适应扩展卡尔曼滤波器平衡交互式多模型中马尔可夫链的模式信任平衡,得到蜂窝网络中相应基站和移动站之间的平滑距离。仿真结果表明,在相同程度的非视距效应的环境下此算法具有良好的定位精度和稳健性,相较于基本交互式多模型算法其定位精度可提升15%以上,且随着非视距效应程度加深,其定位精度可提升至25%。Aiming at the non-line-df-sight phenomenon under the propagation process of distance signals in the study of wireless sensor networks,it is assumed that line-of-sight and non-line-of-sight channel switching in wireless networks is a two-state Markov process,and an interacting multi-model algorithm based on robust adaptation is poposed.Based on the time of anival and signal attenuation values of the wireless sensor network,the robust adaptive extended Kalman filher is used to balance the mode trust balance of the Markov chain in the interacting multi-model.It obtains the smooth distance between the oresponding base station and the mobile station in the wireless sengor network.The simulation results show that this algorithm has good positioning accuracy and robustness under the same degree of non-line-of-sight efet.Compared with the lassical interacting muli-model algorithm,its localzation accuracy can be improved by more than 15%,and can be increased to 25%as the no-line-of-sight fft deepens.

关 键 词：无线传感器网络 无线定位 非视距传播 抗差估计 自适应扩展卡尔曼滤波器 交互式多模型 

分 类 号：TN957[电子电信—信号与信息处理]