Abstract: Accurate short-term traffic prediction is essential for improving the efficiency of data transmission in low Earth orbit (LEO) satellite networks. However, traffic values may be missing due to collector failures, transmission errors, and memory failures in complex space environments. Incomplete traffic time series prevent the efficient utilization of data, which can significantly reduce the traffic prediction accuracy. To overcome this problem, we propose a novel spatio-temporal correlation-based incomplete time-series traffic prediction (ITP-ST) model, which consists of two phases: reconstituting incomplete time series by missing data imputation and making traffic prediction based on the reconstructed time series. In the first phase, we propose a novel missing data imputation model based on the improved denoising autoencoder (IDAE-MDI). Specifically, we combine DAE with the Gramian angular summation field (GASF) to establish the temporal correlation between different time intervals and extract the structural patterns from the time series. Taking advantage of the unique spatio-temporal correlation of the LEO satellite network traffic, we focus on improving the missing data initialization method for DAE. In the second phase, we propose a traffic prediction model based on a multi-channel attention convolutional neural network (TP-CACNN) by combining the spatio-temporally correlated traffic of the LEO satellite network. Finally, to achieve the ideal structure of these models, we use the multi-verse optimizer (MVO) algorithm to select the optimal combination of model parameters. Experiments show that the ITP-ST model outperforms the baseline models in terms of traffic prediction accuracy at different data missing rates, which demonstrates the effectiveness of our proposed model.

低轨卫星网络中基于时空相关性的不完全时间序列流量预测

[1]Appala Naidu T, Raj Arya S, Maurya R, 2019. Dynamic voltage restorer with quasi-Newton filter-based control algorithm and optimized values of PI regulator gains. IEEE J Emerg Sel Top Power Electron, 7(4):2476-2485.

[2]Baggag A, Abbar S, Sharma A, et al., 2021. Learning spatiotemporal latent factors of traffic via regularized tensor factorization: imputing missing values and forecasting. IEEE Trans Knowl Data Eng, 33(6):2573-2587.

[3]Che ZP, Purushotham S, Cho K, et al., 2018. Recurrent neural networks for multivariate time series with missing values. Sci Rep, 8(18):6085.

[4]Coscia M, 2021. Pearson correlations on complex networks. J Compl Netw, 9(6):cnab036.

[5]Deng BW, Xu TW, Yan MD, 2023. UWB NLOS identification and mitigation based on Gramian angular field and parallel deep learning model. IEEE Sens J, 23(22):28513-28525.

[6]Huang J, Luo K, Cao LB, et al., 2022. Learning multiaspect traffic couplings by multirelational graph attention networks for traffic prediction. IEEE Trans Intell Transp Syst, 23(11):20681-20695.

[7]Jiang B, Yan YC, You L, et al., 2023. Robust secure transmission for satellite communications. IEEE Trans Aerosp Electron Syst, 59(2):1598-1612.

[8]Ke RM, Li W, Cui ZY, et al., 2020. Two-stream multi-channel convolutional neural network for multi-lane traffic speed prediction considering traffic volume impact. Transp Res Rec, 2674(4):459-470.

[9]Kumar P, Garg S, Singh A, et al., 2018. MVO-based 2-D path planning scheme for providing quality of service in UAV environment. IEEE Int Things J, 5(3):1698-1707.

[10]Lee WK, Seo HJ, Seo SC, et al., 2022. Efficient implementation of AES-CTR and AES-ECB on GPUs with applications for high-speed FrodoKEM and exhaustive key search. IEEE Trans Circ Syst II Expr Briefs, 69(6):2962-2966.

[11]Li FY, Shang CJ, Li Y, et al., 2020. Interpolation with just two nearest neighboring weighted fuzzy rules. IEEE Trans Fuzzy Syst, 28(9):2255-2262.

[12]Liu W, Ren C, Xu Y, 2021. PV generation forecasting with missing input data: a super-resolution perception approach. IEEE Trans Sustain Energy, 12(2):1493-1496.

[13]Liu ZL, Li X, 2018. Short-term traffic forecasting based on principal component analysis and a generalized regression neural network for satellite networks. J China Univ Posts Telecommun, 25(1):15-28.

[14]Lu BL, Liu ZH, Wei HL, et al., 2021. A deep adversarial learning prognostics model for remaining useful life prediction of rolling bearing. IEEE Trans Artif Intell, 2(4):329-340.

[15]Luo YH, Zhang Y, Cai XR, et al., 2019. E²GAN: end-to-end generative adversarial network for multivariate time series imputation. Proc 28^th Int Joint Conf on Artificial Intelligence, p.3094-3100.

[16]Lv ZL, Peng LH, Cao YJ, et al., 2023. Weak fault feature extraction method of rolling bearings based on MVO-MOMEDA under strong noise interference. IEEE Sens J, 23(14):15732-15740.

[17]Ma Q, Lee WC, Fu TY, et al., 2020. MIDIA: exploring denoising autoencoders for missing data imputation. Data Min Knowl Disc, 34:1859-1897.

[18]Ma XL, Tao ZM, Wang YH, et al., 2015. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data. Transp Res Part C Emerg Technol, 54:187-197.

[19]Ma XL, Dai Z, He ZB, et al., 2017. Learning traffic as images: a deep convolutional neural network for large-scale transportation network speed prediction. Sensors, 17(4):818.

[20]Marchang N, Tripathi R, 2021. KNN-ST: exploiting spatio-temporal correlation for missing data inference in environmental crowd sensing. IEEE Sens J, 21(3):3429-3436.

[21]Miao XY, Wu YY, Chen L, et al., 2023. An experimental survey of missing data imputation algorithms. IEEE Trans Knowl Data Eng, 35(7):6630-6650.

[22]Mukhopadhyay S, Mukherjee A, 2020. ImdLMS: an imputation based LMS algorithm for linear system identification with missing input data. IEEE Trans Signal Process, 68:2370-2385.

[23]Na ZY, Liu Y, Cui Y, et al., 2015. Research on aggregation and propagation of self-similar traffic in satellite network. Int J Hybr Inform Technol, 8:325-338.

[24]Nguyen HD, Vu TL, Slotine JJ, et al., 2021. Contraction analysis of nonlinear DAE systems. IEEE Trans Autom Contr, 66(1):429-436.

[25]Pan ZF, Wang YL, Wang K, et al., 2023. Imputation of missing values in time series using an adaptive-learned median-filled deep autoencoder. IEEE Trans Cybern, 53(2):695-706.

[26]Shen LF, Ma QL, Li S, 2018. End-to-end time series imputation via residual short paths. Proc 10^th Asian Conf on Machine Learning Research, p.248-263.

[27]Su T, Liu YB, Zhao JB, et al., 2021. Probabilistic stacked denoising autoencoder for power system transient stability prediction with wind farms. IEEE Trans Power Syst, 36(4):3786-3789.

[28]Tang D, Wang SY, Liu BR, et al., 2023. GASF-IPP: detection and mitigation of LDoS attack in SDN. IEEE Trans Serv Comput, 16(5):3373-3384.

[29]Tao HM, Deng QQ, Xiao SZ, 2020. Reconstruction of time series with missing value using 2D representation-based denoising autoencoder. J Syst Eng Electron, 31(6):1087-1096.

[30]Tao XL, Liu ZY, Zhao F, et al., 2023. An SSA-LC-DAE method for extracting network security elements. IEEE Trans Netw Sci Eng, 10(2):1175-1185.

[31]Tasdemir Y, Kolay E, Kayabali K, 2013. Comparison of three artificial neural network approaches for estimating of slake durability index. Environ Earth Sci, 68:23-31.

[32]Wang A, Ye YC, Song XZ, et al., 2023. Traffic prediction with missing data: a multi-task learning approach. IEEE Trans Intell Transp Syst, 24(4):4189-4202.

[33]Wang HY, Zhao JP, Su YS, et al., 2022. scCDG: a method based on DAE and GCN for scRNA-seq data analysis. IEEE/ACM Trans Comput Biol Bioinform, 19(6):3685-3694.

[34]Ye L, Hu SB, Yan TT, et al., 2023. GAF representation of millimeter wave drone RCS and drone classification method based on deep fusion network using ResNet. IEEE Trans Aerosp Electron Syst, 59(1):336-346.

[35]Yoon J, Jordon J, Schaar M, 2018. GAIN: missing data imputation using generative adversarial nets. Proc 35^th Int Conf on Machine Learning, p.5689-5698.

[36]You WB, Ding YH, Yao Y, 2020. Static explosion field reconstruction based on the improved biharmonic spline interpolation. IEEE Sens J, 20(13):7235-7240.

[37]Zhang CH, Yu JJQ, Liu Y, 2019. Spatial-temporal graph attention networks: a deep learning approach for traffic forecasting. IEEE Access, 7:166246-166256.

[38]Zhang JJ, Mu XD, Fang JS, et al., 2019. Time series imputation via integration of revealed information based on the residual shortcut connection. IEEE Access, 7:102397-102405.

[39]Zhao L, Song YJ, Zhang C, et al., 2020. T-GCN: a temporal graph convolutional network for traffic prediction. IEEE Trans Intell Transp Syst, 21(9):3848-3858.

[40]Zhou X, Shi J, Gong K, et al., 2021. A novel quench detection method based on CNN-LSTM model. IEEE Trans Appl Supercond, 31(5):4702105.