融合表征转换与模式回归的航迹插补方法

陶冶; 汤锦辉; 闫震; 周臣; 王冲

doi:10.7527/S1000-6893.2025.32106

航空学报 >

0 1 - 0

DOI: https://doi.org/10.7527/S1000-6893.2025.32106

融合表征转换与模式回归的航迹插补方法

陶冶 ,
汤锦辉 ,
闫震 ,
周臣 ,
王冲

展开

1. 空中交通管理系统全国重点实验室
2. 空军装备研究院雷达所
3. 国家空域技术重点实验室

收稿日期: 2025-04-11

修回日期: 2025-08-19

网络出版日期: 2025-08-28

基金资助

军内科研青年创新基金

收起

A trajectory imputation method integrating representation transformation and pattern re-gression

TAO Ye ,
TANG Jin-Hui ,
YAN Zhen ,
ZHOU Chen ,
WANG Chong

Expand

Received date: 2025-04-11

Revised date: 2025-08-19

Online published: 2025-08-28

Fold

摘要

空中交通管制监视系统采集的民航航迹数据存在完整性缺失的问题，不仅影响了空管监视的时空连续性，更制约了基于航迹运行框架下智能化技术的应用潜力。针对这一问题，提出一种融合表征转换与模式回归的航迹插补方法。本方法的技术实现路径包含三个核心构建：首先，构建基于时空四维航迹数据向RGB通道彩色图像表征形式转换的编码模型，通过将缺失航迹插补问题转换为图像修复任务，实现问题域的创新性转换；其次，设计具备多尺度感知能力的回归模型架构，并在各分辨率信息传递路径中分别引入多卷积融合模块与改进型混合注意力机制模块，有效提升复杂时空模式的表征能力；最终，采用提出的基于自适应权重分配的全局与局部特征融合训练策略，系统性地优化模型整体拟合性能。实验环节在真实航迹数据集上构建多维评估体系，结果表明：相较于传统统计插值方法与现有深度学习方法，提出方法针对90%缺失率以下的航迹数据插补精度有明显提升，同时消融实验证实各创新模块与训练策略的有效性。上述结果充分说明提出方法在解决航迹完整性不足问题时，展现出泛化性能优势与工程应用鲁棒性。

关键词： 空中交通管制; 深度学习; 飞行航迹插补; 表征转换; 模式回归; 多尺度回归模型

本文引用格式

陶冶 , 汤锦辉 , 闫震 , 周臣 , 王冲 . 融合表征转换与模式回归的航迹插补方法[J]. 航空学报, 0 : 1 -0 . DOI: 10.7527/S1000-6893.2025.32106

Abstract

Given the challenges posed by incomplete civil aviation trajectory data collected through air traffic control surveillance systems, which compromise spatial-temporal continuity and limit the application potential of trajectory-based intelligent operational frameworks, a novel trajectory imputation method is proposed, integrating representation transformation with pattern regression mecha-nisms. The technical implementation comprises three core components: First, a coding model is developed to transform four-dimensional spatial-temporal trajectories into RGB-channel color image representations. This approach reformulates the missing trajectory completion problem as an image inpainting task, thereby achieving an innovative domain transformation. Second, a multi-scale perception regression model architecture is designed, incorporating multi-kernel fusion modules and improved hybrid attention mechanisms across resolution-specific information pathways. Within each resolution’s information transmission path, multi-kernel fusion modules and hybrid attention mechanisms are integrated to strengthen the model’s ability to represent complex spatial-temporal patterns. Finally, an adaptive weight allocation training strategy is employed to systematically optimizing the model’s fitting performance. During experimental validation, a multi-dimensional evaluation framework is implemented using a real-world trajectory dataset. Results demonstrate that, compared to traditional statistical interpolation and existing deep learning methods, the proposed method achieves superior imputation accuracy for trajectories with missing rates below 90%. Ablation study further confirms the significant contributions of each module and training strategy to overall performance. These findings underscore the method’s robust generalization capabilities and engineering applicability in addressing trajectory integrity challenges.

Key words： air traffic control; deep learning; flight trajectory imputation; representation transformation; pattern regression; multi-scale regression model

参考文献

[1] 陈志杰,汤锦辉,王冲,等.人工智能赋能空域系统，提升空域分层治理能力[J].航空学报,2021,42(4):7-15.
CHEN Z J, TANG J H, WANG C, et al. Using artificial intelligence airspace system to improve airspace hierar-chical governance capability[J]. Acta Aeronautica et Astro-nautica Sinica, 2024, 42(4):7-15(in Chinese).
[2] 陈雨童,胡明华,杨磊,等. 受限航路空域自主航迹规划与冲突管理技术[J].航空学报,2020,41(9):253-270.
CHEN Y T, HU M H, YANG L, et al. Autonomous trajec-tory planning and conflict management technology in re-stricted airspace[J]. Acta Aeronautica et Astronautica Sinica, 2020, 41(9): 253-270 (in Chinese).
[3] 崔亚奇,熊伟,何友.不确定航迹自适应预测模型[J].航空学报,2019,40(5):241-250.
CUI Y Q, XIONG W, HE Y. Adaptive forecast model for uncertain track[J]. Acta Aeronautica et Astronautica Sinica, 2019, 40(5): 241-250 (in Chinese).
[4] 王飞,韩翔宇.基于分形插值的空中交通流量短期预测[J].航空学报,2022,43(9):513-520.
WANG F, HAN X Y. Short-term prediction of air traffic flow based on fractal interpolation[J]. Acta Aeronautica et Astronautica Sinica, 2022, 43(9): 513-520 (in Chinese).
[5] 隋东,蔡向嵘.智能飞行冲突解脱算法的持续学习机制[J/OL].系统工程与电子技术,(2024-04-11)[2024-11-26].https://link.cnki.net/urlid/11.2422.TN.20241126.0955.010.
SUI D, CAI X R. Continual learning mechanism for intelli-gent flight conflict resolution algorithm[J/OL]. Systems Engineering and Electronics (2024-04-11) [2024-11-26]. https://link.cnki.net/urlid/11.2422.TN.20241126.0955.010.
[6] ZHANG Z, GUO D Y, ZHOU S Z, et al. Flight trajectory prediction enabled by time-frequency wavelet transform[J]. Nat Commun, 2023, 14(1): 1-15.
[7] 霍丹,余付平,沈堤,等.基于深度强化学习的多机冲突解决方法的研究[J/OL].计算机科学,(2024-12-11)[2024-12-11].https://link.cnki.net/urlid/50.1075.TP.20241211.1637.016.
HUO D, YU F P, SHEN D, et al. Research on multi-machine conflict resolution based on deep reinforcement learning[J/OL]. Computer Science(2024-12-11) [2024-12-11].https://link.cnki.net/urlid/50.1075.TP.20241211.1637.016.
[8] 李萌.TBO模式下航空器四维航迹规划技术[D],唐山:华北理工大学,2022:9-17.
LI M. 4D trajectory planning technology of aircraft in TBO mode[D]. Tangshan: North China University of Science and Technology, 2022: 9-17 (in Chinese).
[9] CHANDAR B S, RANGANATHAN P, SEMKE W. Imputing ADS-B/GPS dropouts using machine learning models[C]// proceedings of the 2024 IEEE 14th Annual Computing and Communication Workshop and Conference (CCWC). Singapore: IEEE Computer Society, 2024: 1-10.
[10] 王培晓.时空视图学习支持的城市交通数据缺失补全与短期预测[J].测绘学报,2024,53(10):2037.
WANG P X. Missing imputation and short-term prediction of urban traffic flow based on spatiotemporal view learn-ing[J]. Acta Geodaetica et Catographica Sinica, 2024, 53(10): 2037.
[11] XU Y, BAZARJANI A, CHI H G, et al. Uncovering the missing pattern: unified framework towards trajectory im-putation and prediction[C]// proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver: IEEE Computer Society, 2023: 17-24.
[12] MA Q L, LI S, COTTRELL G W. Adversarial joint-learning recurrent neural network for incomplete time series classi-fication[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(4): 1765-1776.
[13] 赵永梅,董云卫.低秩张量补全的时空交通数据预测[J].交通运输工程学报,2024,24(4):243-258.
ZHAO Y M, DONG Y W. Spatio-temporal traffic data prediction based on low-rank tensor completion[J]. Journal of Traffic and Transportation Engineering, 2024, 24(4): 243-258 (in Chinese).
[14] LIN Y, LI Q, GUO D Y, et al. Tensor completion-based trajectory imputation approach in air traffic control[J]. Aer-osp Sci Technol, 2021, 114: 1-11.
[15] ZHAO J Q, RONG C T, DANG X, et al. QAR Data Imputation Using Generative Adversarial Network with Self-Attention Mechanism[J]. Big Data Min Anal, 2024, 7(1): 12-28.
[16] 胡中华,许昕,陈中.无人机三维航迹非均匀三次B样条平滑算法[J].控制工程,2020,27(7):1259-1266.
HU Z H, XU X, CHEN Z. Non-uniform cubic B-spline curves interpolation for UAV 3-D track smoothing[J]. Control Engineering of China, 2020, 27(7): 1259-1266 (in Chinese).
[17] SCH?FER M, STROHMEIER M, LENDERS V, et al. Bringing up OpenSky: A large-scale ADS-B sensor net-work for research[C]// proceedings of the 13th IEEE/ACM International Symposium on Information Processing in Sensor Networks (IPSN), Berlin: IEEE Computer Society, 2014: 1-12.
[18] ZHANG X L, QIAN B Y, LI Y, et al. Context-aware and time-aware attention-based model for disease risk predic-tion with interpretability[J]. IEEE Trans Knowl Data Eng, 2023, 35(4): 3551-3562.
[19] SCARSELLI F, GORI M, TSOI A C, et al. The graph neural network model[J]. IEEE Trans Neural Netw, 2009, 20(1): 61-80.
[20] 杨燕,沈汪良.多尺度细节增强与分层抑噪的图像去雾算法[J/OL].吉林大学学报(工学版), (2024-05-15) [2024-08-12]. https://doi.org//10.13229/j.cnki.jdxbgxb.20240538.
YANG Y, SHEN W L. Multi-scale detail enhancement and layered noise suppression algorithm for image dehaz-ing[J/OL]. Journal of Jilin University(Engineering and Technology Edition), (2024-05-15) [2024-08-12]. https://doi.org//10.13229/j.cnki.jdxbgxb.20240538.
[21] WU X P, YANG H Y, CHEN H, et al. Long-term 4D trajectory prediction using generative adversarial net-works[J]. Transp Res Pt C-Emerg Technol, 2022, 136: 1-13.
[22] NILSSON J, UNGER J. Swedish civil air traffic control dataset[J]. Data Brief, 2023, 48: 1-11.
[23] 陈蜀喆,王子威,龚彪.基于滑动窗口算法的船舶避碰转向点数据挖掘模型[J].中国航海,2025,48(1):124-131.
CHEN S Z, WANG Z W, GONG B. Data mining model of ship collision avoidance turning points based on sliding window algorithm[J]. Navigation of China, 2025, 48(1): 124-131 (in Chinese).
[24] 郭树强,黄蕊,李卿.改进加权朴素贝叶斯的软件缺陷预测算法[J].控制工程,2021,28(3):600-605.
GUO S Q, HUANG R, LI Q. A software defect prediction algorithm using improved weighted naive Bayesian[J]. Control Engineering of China, 2021, 38(3): 600-605 (in Chinese).
[25] 李强,马超,黄民.基于注意力的多尺度残差卷积网络轴承故障诊断[J/OL].电子测量技术,(2025-03-26)[2025-03-26].https://link.cnki.net/urlid/11.2175.TN.20250326.1101.028.
LI Q, MA C, HUANG M. Attention-based multi-scale re-sidual convolutional network for bearing fault diagno-sis[J/OL]. Electronic Measurement Technology, (2025-03-26) [2025-03-26]. https://link.cnki.net/urlid/11.2175.TN.20250326.1101.028.
[26] TAO Y, TANG J, ZHAO X, et al. Multi-scale network with attention mechanism for underwater image enhance-ment[J]. Neurocomputing, 2024, 595: 127926.
[27] CHEN G, ZHANG G P, YANG Z G, et al. Multi-scale patch-GAN with edge detection for image inpainting[J]. Appl Intell, 2023, 53(4): 3917-3932.
[28] DENG Y, HUI S Q, ZHOU S P, et al. Context adaptive network for image inpainting[J]. IEEE Trans Image Pro-cess, 2023, 32: 6332-6345.
[29] YAO C, XIAO J M, TILLO T, et al. Depth map down-sampling and coding based on synthesized view distor-tion[J]. IEEE Transactions on Multimedia, 2016, 18(10): 2015-2022.
[30] PAN J S, SUN D Q, ZHANG J W, et al. Dual convolu-tional neural networks for low-level vision[J]. International Journal of Computer Vision, 2022, 130(6): 1440-1458.
[31] KHAN S, NASEER M, HAYAT M, et al. Transformers in vision: a survey[J]. ACM Comput Surv, 2022, 54(10S): 1-41.
[32] LI X, WANG W H, HU X L, et al. Selective kernel net-works[C]// proceedings of the 32nd IEEE/CVF Confer-ence on Computer Vision and Pattern Recognition (CVPR), Long Beach: IEEE Computer Society, 2019: 1-32.
[33] HU H G, YU C H, ZHOU Q W, et al. HDConv: hetero-geneous kernel-based dilated convolutions[J]. Neural Net-works, 2024, 179: 1-10.
[34] OYEDOTUN O K, AL ISMAEIL K, AOUADA D. Training very deep neural networks: Rethinking the role of skip connections[J]. Neurocomputing, 2021, 441: 105-117.
[35] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011-2023.
[36] ZENG X, PAN Y, ZHANG H, et al. Unpaired salient object translation via spatial attention prior [J]. Neurocom-puting, 2021, 453: 718-730.
[37] LIU H J, LIU F Q, FAN X Y, et al. Polarized self-attention: towards high-quality pixel-wise mapping[J]. Neurocompu-ting, 2022, 506: 158-167.
[38] ZHU M H, JIAO L C, LIU F, et al. Residual spectral-spatial attention network for hyperspectral image classifica-tion[J]. IEEE Trans Geosci Remote Sensing, 2021, 59(1): 449-462.
[39] WANG N, ZHANG Y P, ZHANG L F. Dynamic selection network for image inpainting[J]. IEEE Trans Image Pro-cess, 2021, 30: 1784-1798.
[40] 王超学,王磊.融合多尺度特征和双分支并行的肺结节图像分割网络[J].计算机系统应用,2025,34(4):166-174.
WANG C X, WANG L. Pulmonary nodule image segmen-tation network integrating multi-scale features and dual-branch parallelism[J].Computer Systems Applications, 2025, 34(4): 166-174 (in Chinese).
[41] 申秋慧,张宏军,徐有为,等.知识图谱嵌入模型中的损失函数研究综述[J].计算机科学,2023,50(04):149-158.
SHEN Q H, ZHANG H J, XU Y W, et al. Comprehensive survey of loss functions in knowledge graph embedding models[J]. Computer Science, 2023, 50(4): 149-158 (in Chinese).
[42] LI C, ANWAR S, HOU J, et al. Underwater Image En-hancement via Medium Transmission-Guided Multi-Color Space Embedding [J]. IEEE Trans Image Process, 2021, 30: 4985-5000.
[43] LI Y, FANG A Q, GUO Y M, et al. Smooth fusion of multi-spectral images via total variation minimization for traffic scene semantic segmentation [J]. Engineering Appli-cations of Artificial Intelligence, 2024, 130: 15.
[44] 徐雯,于瓅.基于迭代收缩阈值与深度学习的压缩感知图像重构网络[J].计算机工程与科学,2025,47(3):485-493.
XU W, YU L. A compressive sensing image reconstruction network based on iterative shrinkage thresholding and deep learning[J]. Computer Engineering and Science, 2025, 47(3): 485-493.
[45] 王成,刘坤,杜砾.全参考图像质量指标评价分析[J].现代电子技术,2023,46(21):39-43.
WANG C, LIU K, DU L. Evaluation and analysis of full reference image quality indicators[J]. Modern Electronics Technique, 2023, 46(21): 39-43 (in Chinese).
[46] 吴济洲,张红敏.一种航迹数据高维特征矩阵提取方法[J].指挥控制与仿真,2022,44(1):51-57.
WU J Z, ZHANG H M. Method for extracting high-dimensional feature matrix of track data[J]. Command Control and Simulation, 2022, 44(1): 51-57 (in Chinese).
[47] 孟庆成,万达,吴浩杰,等.基于卷积神经网络的混凝土裂缝图像识别方法[J].沈阳建筑大学学报(自然科学版),2021,37(5):832-840.
MENG Q C, WAN D, WU H J, et al. Image recognition method of concrete cracks based on convolutional neural network[J]. Journal of Shenyang Jianzhu University (Nat-ural Science), 2021, 37(5): 832-840 (in Chinese).

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献