基于BiLSTM模型的BDS-3短期钟差预报精度研究

doi:10.11947/j.AGCS.2024.20230082

测绘学报 ›› 2024, Vol. 53 ›› Issue (1): 65-78.doi: 10.11947/j.AGCS.2024.20230082

基于BiLSTM模型的BDS-3短期钟差预报精度研究

潘雄¹, 黄伟凯¹, 赵万卓¹, 张思莹¹, 张龙杰¹, 金丽宏²

1. 武汉纺织大学计算机与人工智能学院, 湖北武汉 430200;
2. 武汉纺织大学数理科学学院, 湖北武汉 430200

收稿日期:2023-04-07 修回日期:2023-11-27 发布日期:2024-02-06
通讯作者: 金丽宏 E-mail:2022018@wtu.edu.cn
作者简介:潘雄(1973-),男,博士,教授,研究方向为深度学习、卫星导航定位。E-mail:pxjlh@163.com
基金资助:
国家自然科学基金(42174010;41874009);湖北省自然科学基金(2023AFB435)

Research on short-term prediction accuracy of BDS-3 clock bias based on BiLSTM model

PAN Xiong¹, HUANG Weikai¹, ZHAO Wanzhuo¹, ZHANG Siying¹, ZHANG Longjie¹, JIN Lihong²

1. School of Computer Science and Artificial Intelligence, Wuhan Textile University, Wuhan 430200, China;
2. School of Mathematical and Physical Sciences, Wuhan Textile University, Wuhan 430200, China

Received:2023-04-07 Revised:2023-11-27 Published:2024-02-06
Supported by:
The National Natural Science Foundation of China(Nos. 42174010; 41874009); The Hubei Province Natural Science Foundation(No. 2023AFB435)

摘要/Abstract

摘要： 提出了一种改进的北斗钟差预测模型,将传统的单向长短期记忆神经网络(LSTM)扩展为双向长短期记忆网络(BiLSTM),引入了3种自适应匹配超参数的算法提高钟差数据短期预报的精度。首先,对LSTM进行优化,建立BiLSTM模型,介绍了超参数的3种选择方案(粒子群搜索(PSO)、麻雀搜索(SSA)和贝叶斯搜索(BOA)),并给出了相应的适用范围。然后,详细介绍基于超参数优化BiLSTM模型的钟差预报的步骤。最后,利用GFZ卫星钟差数据,从不同轨道类型、5 min采样间隔、15 min采样间隔等方面进行了1、6和12 h的单天和多天预报对比试验,并进行了相应模型的时间复杂度分析。试验结果表明,采用超参数方案优化后的BiLSTM模型在进行1、6和12 h预报时,相较于二次多项式模型、灰色模型、长短期记忆神经网络的模型和BiLSTM模型,平均精度可分别提升86.21%、83.32%、69.99%和55.17%。在3种优化方案中,使用PSO算法对IGSO类型卫星的优化效果较好;使用BOA算法对MEO类型卫星的钟差优化效果较好;使用SSA算法在整体上优化效果最好。虽然经过超参数优化后的BiLSTM模型训练时间相对常用模型较长,但预报速度较快,总体上能够满足实时预报时间要求。

关键词: 钟差预报, BiLSTM, 超参数优化, 神经网络

Abstract: This paper proposes an enhanced model for BeiDou clock bias prediction, which extends the traditional LSTM to BiLSTM and introduces three adaptive hyperparameter matching algorithms (PSO, SSA, BOA) to improve short-term forecast accuracy. Firstly, LSTM is optimized to establish BiLSTM model, and three hyperparameter options have corresponding application scopes. Secondly, detailed steps for prediction with the hyperparameter-optimized BiLSTM model are outlined. Finally, the comparative experiments which consider the factor of the orbit types and sample intervals, are conducted for 1, 6, and 12 hours with the clock products provided by GFZ. The results show that the hyperparameter-optimized BiLSTM outperforms QP, GM, LSTM, and traditional BiLSTM models with an average improvement of 86.21%, 83.32%, 69.99%, and 55.17%. As for the three optimization schemes, SSA exhibits the best overall optimization, and PSO and BOA are more suitable for the IGSO and MEO satellites, respectively. Although the hyperparameter-optimized BiLSTM model takes a long time to train, its rapid forecasting speed can be guaranteed for the requirement of real-time applications.

Key words: clock bias prediction, BiLSTM, hyperparameter optimization, neural network

中图分类号:

P258

潘雄, 黄伟凯, 赵万卓, 张思莹, 张龙杰, 金丽宏. 基于BiLSTM模型的BDS-3短期钟差预报精度研究[J]. 测绘学报, 2024, 53(1): 65-78.

PAN Xiong, HUANG Weikai, ZHAO Wanzhuo, ZHANG Siying, ZHANG Longjie, JIN Lihong. Research on short-term prediction accuracy of BDS-3 clock bias based on BiLSTM model[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(1): 65-78.

参考文献

[1] 李文涛, 颜雄, 夏磊, 等. BDS卫星钟差半参数平差模型异常数据探测与处理[J]. 测绘学报, 2020, 49(1): 55-64. DOI: 10.11947/j.AGCS.2020.20180384. LI Wentao, YAN Xiong, XIA Lei, et al. Abnormal data detection and process by using BDS satellite offset semiparametric adjustment model[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(1): 55-64. DOI: 10.11947/j.AGCS.2020.20180384.
[2] 吕栋, 欧吉坤, 于胜文. 基于MEA-BP神经网络的卫星钟差预报[J]. 测绘学报, 2020, 49(8): 993-1003. DOI: 10.11947/j.AGCS.2020.20200002. LÜ Dong, OU Jikun, YU Shengwen. Prediction of the satellite clock bias based on MEA-BP neural network[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(8): 993-1003. DOI: 10.11947/j.AGCS.2020.20200002.
[3] 王宇谱. GNSS星载原子钟性能分析与卫星钟差建模预报研究[J]. 测绘学报, 2018, 47(7): 1026. DOI: 10.11947/j.AGCS.2018.20170467. WANG Yupu. Research on modeling and prediction of the satellite clock bias and performance evaluation of GNSS satellite clocks[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(7): 1026. DOI: 10.11947/j.AGCS.2018.20170467.
[4] 毛亚, 王潜心, 胡超, 等. BDS-3卫星钟差特性分析[J]. 武汉大学学报(信息科学版), 2020, 45(1): 53-61. MAO Ya, WANG Qianxin, HU Chao, et al. Analysis of the characterization for BDS-3 satellite clock error[J]. Geomatics and Information Science of Wuhan University, 2020, 45(1): 53-61.
[5] TAN Xiaorong, XU Jiangning, LI Fangneng, et al. A new GM (1, 1) model suitable for short-term prediction of satellite clock bias[J]. IET Radar, Sonar & Navigation, 2022, 16(12): 2040-2052.
[6] 李成龙, 陈西宏, 刘继业, 等. 利用自适应TS-IPSO优化的灰色系统预报卫星钟差[J]. 武汉大学学报(信息科学版), 2018, 43(6): 854-859. LI Chenglong, CHEN Xihong, LIU Jiye, et al. Predicting satellite clock errors using grey model optimized by adaptive TS-IPSO[J]. Geomatics and Information Science of Wuhan University, 2018, 43(6): 854-859.
[7] 马朝忠, 朱建青, 韩松辉. 基于ARIMA模型的卫星钟差异常值探测的模型选择方法[J]. 武汉大学学报(信息科学版), 2020, 45(2): 167-172. MA Chaozhong, ZHU Jianqing, HAN Songhui. Model selection method based on ARIMA model in outliers detection of satel-lite clock offset[J]. Geomatics and Information Science of Wuhan University, 2020, 45(2): 167-172.
[8] ZHANG Guochao, HAN Songhui, YE Jun, et al. A method for precisely predicting satellite clock bias based on robust fitting of ARMA models[J]. GPS Solutions, 2021, 26(1): 3.
[9] 潘雄, 杨玉锋, 欧吉坤, 等. BDS-2卫星钟设计年限末期阶段性能评估与分析[J]. 大地测量与地球动力学, 2020, 40(10): 991-999, 1033. PAN Xiong, YANG Yufeng, OU Jikun, et al. Performance evaluation and analysis of the final stage of BDS-2 satellite clock design period[J]. Journal of Geodesy and Geodynamics, 2020, 40(10): 991-999, 1033.
[10] HUANG Guanwen, CUI Bobin, ZHANG Qin, et al. An improved predicted model for BDS ultra-rapid satellite clock off-sets[J]. Remote Sensing, 2018, 10(2): 60.
[11] 朱祥维, 肖华, 雍少为, 等. 卫星钟差预报的Kalman算法及其性能分析[J]. 宇航学报, 2008, 29(3): 966-970, 1052. ZHU Xiangwei, XIAO Hua, YONG Shaowei, et al. The Kalman algorithm used for satellite clock offset prediction and its performance analysis[J]. Journal of Astronautics, 2008, 29(3): 966-970, 1052.
[12] YAN Xiong, LI Wentao, YANG Yufeng, et al. BDS satellite clock offset prediction based on a semiparametric adjustment model considering model errors[J]. Satellite Navigation, 2020, 1(1): 11.
[13] 杨玉锋, 潘雄, 卿晨昕, 等. 基于半参数均值漂移模型的BDS卫星钟差异常探测与修复[J]. 仪器仪表学报, 2020, 41(8): 47-54. YANG Yufeng, PAN Xiong, QING Chenxin, et al. Detection and repair of outliers in BDS satellite clock offset based on semi-parametric mean drift model[J]. Chinese Journal of Scientific Instrument, 2020, 41(8): 47-54.
[14] 黄观文, 崔博斌, 张勤, 等. 附加周期和神经网络补偿的实时钟差预报模型[J]. 宇航学报, 2018, 39(1): 83-88. HUANG Guanwen, CUI Bobin, ZHANG Qin, et al. Real-time clock offset prediction model with periodic and neural network corrections[J]. Journal of Astronautics, 2018, 39(1): 83-88.
[15] 黄博华, 杨勃航, 李锡瑞, 等. 顾及一次差分数据结构特征的钟差预报模型[J]. 武汉大学学报(信息科学版), 2021, 46(8): 1161-1169. HUANG Bohua, YANG Bohang, LI Xirui, et al. Prediction of navigation satellite clock bias considering structure characteristics of single difference data[J]. Geomatics and Information Science of Wuhan University, 2021, 46(8): 1161-1169.
[16] HUANG Bohua, JI Zengxi, ZHAI Renjian, et al. Clock bias prediction algorithm for navigation satellites based on a supervised learning long short-term memory neural network[J]. GPS Solutions, 2021, 25(2): 80.
[17] WEN Tailai, OU Gang, TANG Xiaomei, et al. A novel long short-term memory predicted algorithm for BDS short-term satellite clock offsets[J]. International Journal of Aerospace Engineering, 2021, 2021: 4066275.
[18] KIM M, KIM J. A short-term prediction method of the IGS RTS clock correction by using LSTM network[J]. Journal of Positioning, Navigation, and Timing, 2019, 8(4): 209-214.
[19] HE Shaofeng, LIU Jiulong, ZHU Xiangwei, et al. Research on modeling and predicting of BDS-3 satellite clock bias using the LSTM neural network model[J]. GPS Solutions, 2023, 27(3): 108.
[20] DE PAULO M C M, MARQUES H A, FEITOSA R Q, et al. New encoder-decoder convolutional LSTM neural network architectures for next-day global ionosphere maps forecast[J]. GPS Solutions, 2023, 27(2): 95.
[21] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780.
[22] GERS F A, SCHMIDHUBER J, CUMMINS F. Learning to forget: continual prediction with LSTM[J]. Neural Computa-tion, 2000, 12(10): 2451-2471.

基于BiLSTM模型的BDS-3短期钟差预报精度研究

Research on short-term prediction accuracy of BDS-3 clock bias based on BiLSTM model

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	孙一帆, 刘冰, 余旭初, 谭熊, 余岸竹. 图像级高光谱影像高分辨率特征网络分类方法[J]. 测绘学报, 2024, 53(1): 50-64.
[2]	任加新, 刘万增, 陈军, 张蓝, 陶远, 朱秀丽, 赵婷婷, 李然, 翟曦, 王海清, 周晓光, 侯东阳, 王勇. 知识引导的碎片化栅格地形图比例尺智能识别[J]. 测绘学报, 2024, 53(1): 146-157.
[3]	肖天元, 艾廷华, 余华飞, 杨敏, 刘鹏程. 地图综合图卷积神经网络点群简化方法[J]. 测绘学报, 2024, 53(1): 158-172.
[4]	张云菲, 邱泽航. 利用多层次网眼特征和VAE-PNN模型识别城市道路格网模式[J]. 测绘学报, 2024, 53(1): 189-198.
[5]	姚国标, 黄鹏飞, 龚健雅, 孟飞, 张进. 抗视差的宽基线弱纹理影像自动拼接算法[J]. 测绘学报, 2023, 52(9): 1515-1527.
[6]	安晓亚, 朱余德, 晏雄锋. 卷积神经网络支持下的建筑物选取方法[J]. 测绘学报, 2023, 52(9): 1574-1583.
[7]	何佳星, 郑南山, 丁锐, 张克非, 陈天悦. 粒子群优化卷积神经网络GNSS-IR土壤湿度反演方法[J]. 测绘学报, 2023, 52(8): 1286-1297.
[8]	张蒙蒙, 李伟, 刘欢, 赵旭东, 陶然. 基于形态变换与空间逻辑聚合的高光谱森林树种分类[J]. 测绘学报, 2023, 52(7): 1202-1211.
[9]	宋佳璇, 范大昭, 董杨, 纪松, 李东子. 神经网络学习与灰度信息结合的跨视角影像线特征匹配算法[J]. 测绘学报, 2023, 52(6): 990-999.
[10]	顾小虎, 李正军, 缪健豪, 李星华, 沈焕锋. 高分遥感影像双通道并行混合卷积分类方法[J]. 测绘学报, 2023, 52(5): 798-807.
[11]	赵冰冰, 谭骁勇, 杨学习, 石岩, 邓敏. 融合地理条件驱动效应和图卷积的土地利用演化模拟CA模型[J]. 测绘学报, 2023, 52(5): 831-842.
[12]	余东行, 徐青, 赵传, 郭海涛, 卢俊, 林雨准, 刘相云. 注意力引导特征融合与联合学习的遥感影像场景分类[J]. 测绘学报, 2023, 52(4): 624-637.
[13]	张永显, 马国锐, 崔志祥, 张志军. 面向大视角差的无人机影像序列学习型特征匹配[J]. 测绘学报, 2023, 52(2): 230-243.
[14]	孙根云, 王鑫, 安娜, 张爱竹. 基于多源遥感的大尺度高分辨率不透水面深度学习提取方法[J]. 测绘学报, 2023, 52(2): 272-282.
[15]	禹文豪, 魏铖, 陈佳鑫. 面向不平衡POI类别分布的电子地图多层次标签预测模型[J]. 测绘学报, 2023, 52(2): 318-328.