基于SE-TCN的一维低采样卫星帆板温度遥测数据插补方法

doi:10.7523/j.ucas.2022.032

中国科学院大学学报 ›› 2023, Vol. 40 ›› Issue (6): 810-820.DOI: 10.7523/j.ucas.2022.032

• 电子信息与计算机科学 • 上一篇下一篇

基于SE-TCN的一维低采样卫星帆板温度遥测数据插补方法

许凯凯, 张锐

中国科学院微小卫星创新研究院, 上海 201203;中国科学院大学, 北京 100049

收稿日期:2021-12-07 修回日期:2022-04-06 发布日期:2022-04-26
通讯作者: 张锐,E-mail：acumen_zhang@163.com
基金资助:
航天系统部委预研基金（2019ZFC1504201）资助

An interpolation method for temperature telemetry data of one-dimensional low-sampling satellite panel based on SE-TCN

XU Kaikai, ZHANG Rui

Innovation Academy for Microsatellites, Chinese Academy of Sciences, Shanghai 201203, China;University of Chinese Academy of Sciences, Beijing 100049, China

Received:2021-12-07 Revised:2022-04-06 Published:2022-04-26

摘要/Abstract

摘要： 针对因入境时间短、组帧错误等原因导致的卫星帆板温度遥测数据缺失问题，提出一种基于引入注意力机制的时间卷积网络（SE-TCN）的自回归预测方法。温度遥测数据可看作是具有较强规律性的渐周期信号，采用SE-TCN对历史数据到未来数据的映射进行拟合完成缺失值的插补，同时为表征对实际缺失数据集的插补效果，增加评价指标的计算方式，有效解决了使用物理模型仿真和统计学方法插值偏差过大，及无法计算实际插值效果的问题。与长短时记忆网络和时间卷积网络等模型相比，SE-TCN在测试集和实际缺失数据集上均得到了更好的插值效果。

关键词: 遥测数据, 时序数据, 缺失值插补, 时间卷积网络, 低采样

Abstract: This paper proposes an autoregressive prediction method based on time convolutional network with attentional mechanism(time convolution network with squeeze and excitation, SE-TCN), to solve the problem of missing telemetry data of satellite panel temperature due to short entry time, framing error, and other reasons. Temperature telemetry data is considered to be a strong regularity of periodic signal, so this paper adopts the SE-TCN model to map from historical data to the data in the future, which completes the missing value interpolation and effectively solves the problem that the interpolation deviation of the physical model and statistical method is too large. At the same time, in order to characterize the interpolation effect on the actual missing data set, the calculation method of the evaluation index is added in this paper. Compared with long short-term memory network and time convolutiion network models, SE-TCN has a better interpolation effect on both the test set and the actual missing data set.

Key words: telemetry data, time series data, missing value interpolation, time convolution network, low sampling

中图分类号:

TP391

许凯凯, 张锐. 基于SE-TCN的一维低采样卫星帆板温度遥测数据插补方法[J]. 中国科学院大学学报, 2023, 40(6): 810-820.

XU Kaikai, ZHANG Rui. An interpolation method for temperature telemetry data of one-dimensional low-sampling satellite panel based on SE-TCN[J]. Journal of University of Chinese Academy of Sciences, 2023, 40(6): 810-820.

参考文献

[1] 王亚坤, 杨凯飞, 张婕, 等. 卫星在轨故障案例与人工智能故障诊断[J]. 中国空间科学技术, 2022, 42(1):16-29. DOI:10.16708/j.cnki.1000-758X.2022.0002.
[2] Lepot M, Aubin J B, Clemens F. Interpolation in time series:an introductive overview of existing methods, their performance criteria and uncertainty assessment[J]. Water, 2017, 9(10):796. DOI:10.3390/w9100796.
[3] Sibson R. A brief description of natural neighbour interpolation[M/OL]//John W & Sons. Interpreting Multivariate Data. In:Barnett, V., Ed. New York, 1981:21-36[2022-03-13]. https://www.scirp.org/(S(i43dyn45teexjx455qlt3d2q))/reference/ReferencesPapers.aspx?ReferenceID=1553046.
[4] Gnauck A. Interpolation and approximation of water quality time series and process identification[J]. Analytical and Bioanalytical Chemistry, 2004, 380(3):484-492. DOI:10.1007/s00216-004-2799-3.
[5] Knotters M, Heuvelink G B M. A disposition of interpolation techniques[R/OL]. Wageningen, Holland:Wettelijke Onderzoekstaken Natuur & Milieu, 2010[2022-03-13]. https://edepot.wur.nl/139504.
[6] Dutta R S C, Minocha S. On the phase interpolation problem:a brief review and some new results[J]. Sadhana, 1991, 16(3):225-239. DOI:10.1007/BF02812044.
[7] Carrizosa E, Olivares N A V, Ramírez C P. Time series interpolation via global optimization of moments fitting[J]. European Journal of Operational Research, 2013, 230(1):97-112. DOI:10.1016/j.ejor.2013.04.008.
[8] Chen Y, Kopp G A, Surry D. Interpolation of wind-induced pressure time series with an artificial neural network[J]. Journal of Wind Engineering and Industrial Aerodynamics, 2002, 90(6):589-615. DOI:10.1016/S0167-6105(02)00155-1.
[9] Kulesh M, Holschneider M, Kurennaya K. Adaptive metrics in the nearest neighbours method[J]. Physica D:Nonlinear Phenomena, 2008, 237(3):283-291. DOI:10.1016/j.physd.2007.08.019.
[10] Lütkepohl H. New Introduction to Multiple Time Series Analysis[M]. Berlin, Heidelberg:Springer Berlin Heidelberg, 2005. DOI:10.1007/978-3-540-27752-1.
[11] 张晓琴, 王敏. 基于主成分分析的成分数据缺失值插补法[J]. 应用概率统计, 2016, 32(1):101-110. DOI:10.3969/j.issn.1001-4268.2016.01.007.
[12] Salakhutdinov R, Mnih A. Bayesian probabilistic matrix factorization using Markov chain Monte Carlo[C]//ICML'08:Proceedings of the 25th international conference on Machine learning. 2008:880-887. DOI:10.1145/1390156.1390267.
[13] Babu C N, Sure P. Partitioning and interpolation based hybrid ARIMA-ANN model for time series forecasting[J]. Sādhanā, 2016, 41(7):695-706. DOI:10.1007/s12046-016-0508-5.
[14] Hundman K, Constantinou V, Laporte C, et al. Detecting spacecraft anomalies using LSTMs and nonparametric dynamic thresholding[C]//KDD'18:Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 2018:387-395. DOI:10.1145/3219819.3219845.
[15] 王少影, 张宇, 孟宪红, 等. 机器学习算法对涡动相关缺失通量数据的插补研究[J]. 高原气象, 2020, 39(6):1348-1360. DOI:10.7522/j.issn.1000-0534.2019.00142.
[16] Raubitzek S, Neubauer T. A fractal interpolation approach to improve neural network predictions for difficult time series data[J]. Expert Systems With Applications, 2021, 169:114474. DOI:10.1016/j.eswa.2020.114474.
[17] Bai S J, Kolter J Z, Koltun V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[EB/OL]. arXiv:1803.01271v2[cs.LG]. (2018-04-19)[2022-03-13]. https://arxiv.org/abs/1803.01271.
[18] Steinmetz C J, Reiss J D. Efficient neural networks for real-time analog audio effect modeling[EB/OL]. arXiv:2102.06200v1[eess.AS]. (2021-02-11)[2022-03-13]. https://arxiv.org/abs/2102.06200.
[19] Alqahtani S, Mishra A, Diab M. Efficient convolutional neural networks for diacritic restoration[EB/OL]. 2019:arXiv:1912.06900[cs.CL]. (2019-12-14)[2022-03-13]. https://arxiv.org/abs/1912.06900.
[20] Dai W, Zhang J S, Gao Y M, et al. Formant tracking using dilated convolutional networks through dense connection with gating mechanism[C]//Interspeech 2020. ISCA:ISCA, 2020:150-154. DOI:10.21437/interspeech.2020-1804.
[21] Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. June 18-23, 2018, Salt Lake City, UT, USA. IEEE, 2018:7132-7141. DOI:10.1109/CVPR.2018.00745.
[22] 何利健, 张锐, 陈文卿. 基于SE-TCN网络模型的太阳电池阵温度异常检测[J]. 上海航天(中英文), 2021, 38(5):8-16. DOI:10.19328/j.cnki.2096-8655.2021.05.002.
[23] Chen W Q, Zhang R, Liu H, et al. A novel method for solar panel temperature determination based on a wavelet neural network and Hammerstein-Wiener model[J]. Advances in Space Research, 2020, 66(8):2035-2046. DOI:10.1016/j.asr.2020.07.002.
[24] Brownlee J. Deep learning for time series forecasting:predict the future with MLPs, CNNs and LSTMs in Python[M/OL]. Machine Learning Mastery, 2018[2022-03-13]. https://machinelearningmastery.com/deep-learning-for-time-series-forecasting/.
[25] Beale M H, Hagan M T, Demuth H B. Deep learning toolbox^TM reference[M/OL]. MATLAB (r) R2021a The MathWorks, Inc, 2021[2022-03-13]. https://ww2.mathworks.cn/help/pdf_doc/deeplearning/index.html.

基于SE-TCN的一维低采样卫星帆板温度遥测数据插补方法

An interpolation method for temperature telemetry data of one-dimensional low-sampling satellite panel based on SE-TCN

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics

本文评价

访问统计

联系我们