Abstract：In order to address the air traffic complexity prediction problem, an ICA-LSTM prediction model is established by constructing the air traffic interdependent network and extracting the nonlinear spatiotemporal dynamic characteristics of air traffic data, which improves the accuracy of prediction. Firstly, based on the complex network theory, taking aircraft and control sectors as the research objects, a flight-control air traffic interdependence network was established. Secondly, network characteristic indexes were selected from the three dimensions of “point-line-surface”, and the common factors of these indexes were extracted using factor analysis method, and an air traffic complexity model was established. Finally, a spatiotemporal series of air traffic data is constructed, the independent component analysis (ICA) is used to extract data sample characteristics, and an ICA-LSTM air traffic complexity prediction model is established. ADS-B operational data from the Beijing terminal area is used for verification, and the results show that the model can effectively predict air traffic complexity. Moreover, compared with traditional LSTM and SVM models, the ICA-LSTM model has higher prediction accuracy.
齐雁楠, 王新彤, 吴祚禹. 基于ICA-LSTM的空中交通相依网络复杂度预测[J]. 复杂系统与复杂性科学, 2025, 22(4): 37-45.
QI Yannan, WANG Xintong, WU Zuoyu. Complexity Prediction of Air Traffic Interdependent Network Based on ICA-LSTM[J]. Complex Systems and Complexity Science, 2025, 22(4): 37-45.
[1] 吴明功,叶泽龙,温祥西, 等. 基于复杂网络的空中交通复杂性识别方法[J]. 北京航空航天大学学报, 2020,46(5): 839850. WU M G,YE Z L,WEN X X, et al. Air traffic complexity recognition method based on complex networks[J]. Journal of Beijing University of Aeronautics and Astronautics,2020,46(5):839-850. [2] 刘寅. 考虑高度层的多层空中交通网络模型构建与分析[D]. 北京:北京交通大学, 2022. LIU Y. Multi-layer air traffic network model construction and analysis considering altitude level[D]. Beijing:Beijing Jiaotong University, 2022. [3] 刘武君,温祥西,范文广, 等. 管制飞行状态相依网络演化过程混沌性研究[J]. 航空工程进展, 2023,14 (4): 137148. LIU W J, WEN R X, FAN W G, et al. Study on chaos during evolution process of control-aircraft state interdependence network[J]. Advances in Aeronautical Science and Engineering, 2023,14 (4): 137148. [4] 李昂,聂党民,温祥西, 等. 基于相依网络和SVM的管制系统运行态势评估[J]. 系统工程与电子技术, 2021,43(5): 12871294. LI A, NIE D M, WEN X X, et al. Evaluation of the operation situation of the control system based on dependent network and SVM[J]. Journal of Systems Engineering and Electronics, 2021,43(5): 12871294. [5] 李昂,聂党民,温祥西,等. 管制飞行状态相依网络演化过程[J]. 航空学报, 2021,42 (9): 481493. LI A, NIE D M, WEN R X, et al. Evolution process of control-aircraft state interdependent network[J]. Acta Aeronautica et Astronautica Sinica, 2021,42(9): 481493. [6] 姜山,丁治明,徐馨润, 等. 面向路网交通流态势预测的图神经网络模型[J]. 计算机科学与探索, 2021,15(6): 10841091. JIANG S,JING Z M,XU X R, et al. Graph neural network for traffic flow situation prediction[J]. Frontiers of Computer Science and Technology, 2021,15(6): 10841091. [7] 李桂毅,胡明华. 考虑航段相关性的航路拥挤态势多模型融合动态预测方法[J]. 交通运输系统工程与信息, 2018,18(1): 215222. LI G Y, HU M H. Multi-model fusion dynamic prediction method of enroute congestion situation with considering the correlation of air route segment[J]. Journal of Transportation System Engineering and Information Technology, 2018,18(1): 215222. [8] 吕业,邓玉俊,杨慧中. 基于类别特征提取的组合支持向量机模型[J]. 化工学报, 2011,62(8): 21642169. LÜ Y, DENG Y J, YANG H Z. Combinatorial support vector machine model based on feature extraction of categories[J]. CIESC Journal, 2011, 62 (8): 21642169. [9] 何光辉,张太平,唐远炎.非监督子空间学习中关联度量多维尺度分析[J]. 中国图象图形学报, 2011,16(12): 21522158. HE G H, ZHANG T P, TANG Y Y. Correlation-based multidimensional scaling for unsupervised subspace learning[J]. Journal of Image and Graphics, 2011, 16 (12): 21522158. [10] 靳洋. 基于PCA/ICA的图像特征提取算法研究[D]. 西安:西安电子科技大学, 2014. JIN Y. Research on image feature extraction algorithm based on PCA/ICA[D]. Xi’an: Xidian University, 2014. [11] 崔建国,碧嘉,蒋丽英, 等. 基于ICA与深度学习的发动机气路系统状态监测[J]. 振动测试与诊断, 2020,40(6): 11501155,1234. CUI J G, ZHOU B J, JIANG L Y, et al. Condition monitoring of aero-engine gas path system based on independent component analysis and deep learning[J]. Journal of Vibration, Measurement & Diagnosis, 2020,40(6): 11501155,1234. [12] 鄂建伟. ICA算法及其在时间序列分析中的应用研究[D]. 西安:西安电子科技大学, 2021. E J W. Research onindependent component analysis algorithm and its applications to the time series analysis[D]. Xi’an:Xidian University, 2021. [13] JONATAN H. WERNER K. A combined independent component analysis-neural network model for forecasting exchange rate variation[J]. Applied Soft Computing, 2019, 83(4): 105654. [14] 钟丽君,宾晟,袁敏, 等. 多功能复杂网络模型及其应用[J]. 复杂系统与复杂性科学, 2019,16(2): 3140. ZHONG L J, BIN S, YUAN M, et al. Multifunctional complex network model and application[J]. Complex Systems and Complexity Science, 2019,16(2): 3140. [15] 王红勇,赵嶷飞,温瑞英. 基于复杂网络的空中交通复杂性度量方法[J]. 系统工程, 2014,32(3): 112118. WANG H Y, ZHAO Y F, WEN R Y. Air traffic complexity metrics based on complex network[J]. Systems Engineering, 2014,32(3): 112118. [16] 顾雪松,迟国泰,程鹤. 基于聚类因子分析的科技评价指标体系构建[J]. 科学学研究, 2010,28(4): 508514. GU X S,CHI G T,CHENG H. The establishment of science and technology evaluation indicators system based on cluster-factor analysis[J]. Studies in Science of Scinece, 2010,28(4): 508514. [17] COMON P. Independent component analysis, a new concept[J]. Signal Processing, 1994,36(3): 287314. [18] HYVÄRINEN A. Fast and robust fixed-point algorithms for independent component analysis[J]. IEEE Transactions on Neural Networks, 1999,10(3): 626634.
