Prediction of tropospheric delay based on the LSTM model of Keras platform
-
摘要: 对流层延迟是影响全球卫星导航系统(GNSS)测量精度的重要因素. 针对现有对流层延迟模型稳定性差,精度较低等问题,在无实测气象参数条件下,提出一种基于Keras平台的长短期记忆神经网络(LSTM)的对流层延迟预测模型. 选取全球均匀分布的8个测站,使用其2016年第90-131年积日共42 天的整点对流层延迟数据预测其第132-136年积日的整点数据. 以国际GNSS服务(IGS)中心提供的对流层产品为真值,分析比较LSTM模型和反向传播(BP)神经网络模型的预测效果. 研究表明,LSTM模型预测结果的均方根误差基本达到mm级,其平均绝对误差和平均绝对百分比误差均比BP模型低,LSTM模型在精度和稳定性上较BP模型均有明显提高;LSTM模型在中高纬区域的均方根误差(RMSE)均值达到7.82 mm,中高纬地区更适合使用该模型.Abstract: Tropospheric delay is a vital factor that influence the measurement accuracy of GNSS. To solve the problems of poor stability and low accuracy of existing tropospheric delay model, a prediction model of tropospheric delay based on the Long-Short Term Memory neural network (LSTM) of Keras platform is proposed in the absence of measured meteorological parameters. Eight stations evenly distributed around the world were selected to use their 42-day hourly tropospheric delay data from the 90th to 131st day of 2016 to predict their hourly data of 132nd to 136th day. Based on the troposphere products provided by International GNSS Service (IGS) center, the prediction effects of LSTM model and back propagation neural network (BP) model were analyzed and compared. The result shows that the root mean square error of LSTM model basically reaches mm level, and the mean absolute error and mean absolute percentage error of LSTM model are lower than those of BP model, and the accuracy and stability of LSTM model are significantly improved compared with BP model. LSTM model has an average RMSE of 7.82 mm in mid and high latitude, which shows it is more suitable for mid and high latitude.
-
Key words:
- tropospheric delay /
- Keras /
- long-short term memory neural network /
- time series /
- prediction accuracy
-
[1] 赵志浩. 几种对流层延迟改正模型对GPS精密单点定位结果的影响[D]. 西安:长安大学, 2014. [2] 章红. 半参数模型在香港地区对流层拟合中的应用研究[D]. 成都:成都理工大学, 2018. [3] HU Y F, YAO Y B. A new method for vertical stratification of zenith tropospheric delay[J]. Advances in space research,2019, 63(9):2857-2866.DOI: 10.1016/j.asr.2018.10.035. [4] 丁晓光. 对流层延迟改正在GPS数据处理中的应用与研究[D].西安:长安大学, 2009. [5] ZHENG D Y, HU W S, WANG J, et al. Research on regional zenith tropospheric delay based on neural network technology[J]. Survey review, 2015, 47(343):286-295.DOI: 10.1179/1752270614Y.0000000130. [6] 陈阳, 胡伍生, 严宇翔, 等. 基于神经网络模型误差补偿技术的对流层延迟模型研究[J]. 大地测量与地球动力学, 2018, 38(6):577-580,586. [7] 吕慧珠, 黄文德, 闻德保. 一种基于频谱分析和AR补偿的对流层延迟预报模型[J]. 大地测量与地球动力学, 2015, 35(2):283-286,308. [8] 李伟捷, 刘根友. 小波自回归模型在天顶对流层延迟预测中的应用[J]. 导航定位学报, 2019, 7(2):131-137. [9] 任超, 刘中流, 梁月吉,等. 基于EEMD-SARIMA的对流层延迟预测模型研究[J]. 大地测量与地球动力学, 2018, 38(9):953-957,963. [10] CHEN J P, WANG J G, WANG A H, et al. SHAtropE-A regional Gridded ZTD model for China and the surrounding areas[J]. Remote sensing, 2020, 12(1):165.DOI: 10.3390/rs12010165. [11] 尹为松, 陶庭叶, 邓清军, 等. 遗传算法优化的GPS对流层延迟内插算法[J]. 测绘科学, 2016, 41(1):180-184. [12] 李剑锋, 吴林弟, 胡伍生, 等. 基于BP神经网络算法的对流层湿延迟计算[J]. 东南大学学报(自然科学版), 2013, 43(增刊2):355-359. [13] DAI W J, YAN H N, CHEN B Y, et al. Regional spatio-temporal zenith tropospheric delay modelling using independent component analysis[J]. Survey review, 2019, 51(369):544-552.DOI: 10.1080/00396265.2018.1515812. [14] 彭正耀, 郭智方, 段彬. 基于Keras框架的人工智能医疗废弃物分类的研究[J]. 科技与创新, 2020(10):1-3. [15] HUAN J, LI H, LI M B, et al. Prediction of dissolved oxygen in aquaculture based on gradient boosting decision tree and long short-term memory network:A study of Chang Zhou fishery demonstration base, China[J]. Computers and electronics in agriculture, 2020, 175:455-463.DOI: 10.1016/j.cpmpag.2020.105530. [16] 庄瑞刚. 基于双向循环神经网络的交通流预测算法研究[D].西安:西安电子科技大学, 2019. [17] LI M X, LU F, ZHANG H, et al. Predicting future locations of moving objects with deep fuzzy-LSTM networks[J]. Transportmetrica a:transport science, 2020, 16(1):119-136.DOI: 10.1080/23249935.2018.1552334. [18] 刘新, 赵宁, 郭金运, 等. 基于LSTM神经网络的青藏高原月降水量预测[J]. 地球信息科学学报, 2020, 22(8):1617-1629. [19] 罗亦泳, 张静影, 陈郡怡, 等. 基于相空间重构和高斯过程回归的对流层延迟预测[J]. 武汉大学学报(信息科学版).[2020-06-17].https://doiorg/10.13203/j.whugis20190018. [20] 马停停, 冀天娇, 杨冠羽, 等. 基于长短时记忆神经网络的手足口病发病趋势预测[J]. 计算机应用,2020:1-7.
点击查看大图
计量
- 文章访问数: 416
- HTML全文浏览量: 114
- PDF下载量: 27
- 被引次数: 0