首页|一种LM-BP加速搜索的周跳探测与修复方法

一种LM-BP加速搜索的周跳探测与修复方法

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
针对传统三频周跳探测与修复方法中的不敏感、漏检以及效率较低等问题,提出一种基于莱文伯格-马夸特(LM)-反向传播(BP)神经网络加速搜索法的伪距相位组合与电离层残差组合联合周跳探测与修复方法:利用 2 个伪距相位组合以减少不敏感周跳数量,利用 1 个电离层残差组合以提高小周跳探测敏感度;在构成 3 个线性无关的组合观测值后,使用 LM-BP 加速搜索算法进行周跳探测与修复.实验结果表明,相对常规的伪距相位组合与电离层残差组合联合方法,该方法能够提高周跳探测与修复性能,可探测小至 1 个的周跳,探测与修复整体时效有较大提升.
A cycle slip detection and repair method for LM-BP accelerated search
Aiming at the problems of insensitivity to some cycle slips,missed detections and low efficiency in traditional three-frequency cycle slip detection and repair methods,the paper proposed a joint cycle slip detection and repair method for pseudorange phase combinations and ionospheric residual combinations based on Levenberg-Marquardt(LM)-back propagation(BP)accelerated search algorithm:to minimize the occurrence of insensitive cycle slips,a combination of two sets of pseudorange phase measurements was employed,as well as one ionospheric residual combination was used to increase sensitivity to small cycle slips;and then after forming three linearly independent combination observations,the LM-BP accelerated search algorithm was used for cycle slip detection and repair.Experimental results showed that the proposed method would improve the cycle slip detection and repair performance compared with the conventional combined method of pseudo range phase combination and ionospheric residual error combination,which could detect cycle slips as small as one cycle,and the overall efficiency of detection and repair would be increased.

BeiDou navigation satellite system(BDS)cycle slip detection and repairLevenberg-Marquardt(LM)-back propagation(BP)algorithmneural networkpseudorange carrier phase combinationionospheric residual combination

梁凌峰、李克昭、张捍卫、雷伟伟、岳哲

展开 >

河南理工大学 测绘与国土信息工程学院,河南 焦作 454000

北斗导航应用技术协同创新中心,郑州 450001

北斗卫星导航系统(BDS) 周跳探测与修复 莱文伯格-马夸特(LM)-反向传播(BP)算法 神经网络 伪距载波相位组合 电离层残差组合

国家自然科学基金项目国家自然科学基金项目国家自然科学基金项目

417740394207400241931075

2024

10.16547/j.cnki.10-1096.20240105

导航定位学报

导航定位学报

CSTPCD北大核心
影响因子：0.72
ISSN：
年,卷(期)：2024.12(1)
  • 20