基于自生成深度神经网络的4D航迹预测

doi:10.11772/j.issn.1001-9081.2020081198

计算机应用 ›› 2021, Vol. 41 ›› Issue (5): 1492-1499.DOI: 10.11772/j.issn.1001-9081.2020081198

所属专题：前沿与综合应用

• 应用前沿、交叉与综合 • 上一篇下一篇

基于自生成深度神经网络的4D航迹预测

李旭娟^1,2, 皮建勇^1,2, 黄飞翔³, 贾海朋³

1. 贵州大学计算机科学与技术学院, 贵阳 550025;
2. 贵州大学云计算与物联网研究中心, 贵阳 550025;
3. 中国民用航空西南地区空中交通管理局贵州分局, 贵阳 550005

收稿日期:2020-08-10 修回日期:2020-11-19 出版日期:2021-05-10 发布日期:2020-12-09
通讯作者: 李旭娟
作者简介:李旭娟(1987-),女,山东昌邑人,硕士研究生,主要研究方向:深度学习、数据挖掘;皮建勇(1973-),男,四川广安人,副教授,博士,CCF会员,主要研究方向:分布式计算、复杂系统、人工智能;黄飞翔(1987-),男,江苏大丰人,工程师,主要研究方向:通信、导航及监视;贾海朋(1982-),男,山东青州人,工程师,硕士,主要研究方向:通信与信息系统。

Self-generated deep neural network based 4D trajectory prediction

LI Xujuan^1,2, PI Jianyong^1,2, HUANG Feixiang³, JIA Haipeng³

1. College of Computer Science and Technology, Guizhou University, Guiyang Guizhou 550025, China;
2. Research Center of Cloud Computing and Internet of Things, Guizhou University, Guiyang Guizhou 550025, China;
3. Guizhou Branch of CAAC Southwest Regional Administration, Guiyang Guizhou 550005, China

Received:2020-08-10 Revised:2020-11-19 Online:2021-05-10 Published:2020-12-09

摘要/Abstract

摘要： 针对四维（4D）航迹预测的实时性不强和存在迭代误差的问题，提出了一种自动生成的条件变分自动编码器（AutoCVAE）。它以编码-解码的形式直接对未来一段时间的航迹进行预测，并能灵活选取观测点个数和预测步长。该方法以处理后的广播式自动相关监视（ADS-B）数据为引导，以减小预测误差为目标，通过贝叶斯优化的方法，在预定义的搜索空间内进行模型结构搜索，每一次的超参数取值都会参考之前的评估结果，使得每一次的模型结构都能向目标更靠近一点，最终实现了一个基于ADS-B数据的高精度的4D航迹预测模型。实验得出，所提模型能快速准确地进行航迹的实时预测，其中经纬度平均绝对预测误差（MAE）均小于0.03°，高度MAE小于30 m，各时刻点的时间误差也不会超过10 s，每次批量预测轨迹的延迟时间不超过0.2 s。

关键词: 航迹预测, 条件变分自动编码器, 深度生成模型, 数据挖掘, 广播式自动相关监视

Abstract: Since 4-Dimensional (4D) trajectory prediction is not real-time and has the iterative error, an Automatically generated Conditional Variational Auto-Encoder (AutoCVAE) was proposed. It is in the form of encoding-decoding to predict the future trajectory directly, and can select observation number and prediction step flexibly. The method was guided by the preprocessed Automatic Dependent Surveillance-Broadcast (ADS-B) data, and with the reduction of the prediction error as the goal. By means of Bayesian optimization, the model structure was searched within the predefined search space. The hyper parameter values of each time were chosen by referencing the previous evaluation results, so that the structure of the new model obtained in each time was able to be closer to the target, and ultimately, a high precision 4D trajectory prediction model based on ADS-B data was completed. In the experiments, the proposed model was able to predict the trajectory quickly and accurately in real time with the Mean Absolute Error (MAE) of both latitude and longitude less than 0.03 degrees, the altitude MAE under 30 m, the time error at each time point not exceeded 10 s, and each batch trajectory prediction delay within 0.2 s.

Key words: trajectory prediction, Conditional Variational Auto-Encoder (CVAE), deep generative model, data mining, Automatic Dependent Surveillance-Broadcast (ADS-B)

中图分类号:

TP391.4

李旭娟, 皮建勇, 黄飞翔, 贾海朋. 基于自生成深度神经网络的4D航迹预测[J]. 计算机应用, 2021, 41(5): 1492-1499.

LI Xujuan, PI Jianyong, HUANG Feixiang, JIA Haipeng. Self-generated deep neural network based 4D trajectory prediction[J]. Journal of Computer Applications, 2021, 41(5): 1492-1499.

参考文献

[1] 高海超, 吴嘉慧. 欧美下一代空管系统规划对比分析及启示[J]. 指挥信息系统与技术,2017,8(4):83-87.(GAO H C,WU J H. Comparative analysis and enlightenment of next generation air traffic management system planning in Euramerica[J]. Command Information System and Technology,2017,8(4):83-87.)
[2] 周雄飞, 胡明华. 空域容量评估研究综述[J]. 中国民航飞行学院学报,2016,27(6):37-40.(ZHOU X F,HU M H. An overview on airspace capacity evaluation[J]. Journal of Civil Aviation Flight University of China,2016,27(6):37-40.)
[3] KASABOV N K. Time-Space,Spiking Neural Networks and BrainInspired Artificial Intelligence[M]. Berlin:Springer,2019:9-11.
[4] 吴鹍, 潘薇. 基于数据挖掘的四维飞行轨迹预测模型[J]. 计算机应用,2007,27(11):2637-2639. (WU K,PAN W. 4-D trajectory prediction model based on data mining[J]. Journal of Computer Applications,2007,27(11):2637-2639.)
[5] 马兰, 高永胜. 基于ADS-B数据挖掘的4D航迹预测方法[J]. 中国民航大学学报,2019,37(4):1-4.(MA L,GAO Y S. Fourdimensional trajectory prediction method based on ADS-B data mining[J]. Journal of Civil Aviation University of China,2019,37(4):1-4.)
[6] 林毅, 张建伟, 武喜萍, 等. 基于GMM的航班轨迹预测算法研究[J]. 工程科学与技术,2018,50(4):104-109.(LIN Y,ZHANG J W, WU X P, et al. Study on algorithm for flight trajectory prediction based on GMM[J]. Advanced Engineering Sciences, 2018,50(4):104-109.)
[7] WU Z,TIAN S,MA L. A 4D trajectory prediction model based on the BP neural network[J]. Journal of Intelligent Systems,2020,29(1):1545-1557.
[8] 王涛波, 黄宝军. 基于改进卡尔曼滤波的四维飞行航迹预测模型[J]. 计算机应用,2014,34(6):1812-1815.(WANG T B, HUANG B J. 4D flight trajectory prediction model based on improved Kalman filter[J]. Journal of Computer Applications, 2014,34(6):1812-1815.)
[9] 韩剑峰. 一种基于雷达数据融合的航班4D航迹预测方法[J]. 软件工程,2019,22(9):8-11.(HAN J F. A flight 4D trajectory prediction method based on radar data fusion[J]. Software Engineering,2019,22(9):8-11.)
[10] GOMEZ-GONZALEZ S,PROKUDIN S,SCHÖLKOPF B,et al. Real time trajectory prediction using deep conditional generative models[J]. IEEE Robotics and Automation Letters,2020,5(2):970-976.
[11] 刘莹. 基于深度学习的轨迹预测[D]. 成都:电子科技大学, 2019:27-31. (LIU Y. Trajectory prediction based on deep learning[D]. Chengdu:University of Electronic Science and Technology of China,2019:27-31.)
[12] 权波, 杨博辰, 胡可奇, 等. 基于LSTM的船舶航迹预测模型[J]. 计算机科学,2018,45(11A):126-131.(QUAN B,YANG B C,HU K Q,et al. Prediction model of ship trajectory based on LSTM[J]. Computer Science,2018,45(11A):126-131.)
[13] YAO D,ZHANG C,ZHU Z,et al. Trajectory clustering via deep representation learning[C]//Proceedings of the 2017 International Joint Conference on Neural Networks. Piscataway:IEEE,2017:3880-3887.
[14] SOHN K, LEE H, YAN X C. Learning structured output representation using deep conditional generative models[C]//Proceedings of the 28th International Conference on Neural Information Processing Systems. Cambridge:MIT Press,2015:3483-3491.
[15] ZHANG X,ZHANG J,WU S,et al. Aircraft monitoring by the fusion of satellite and ground ADS-B data[J]. Acta Astronautica, 2018,143:398-405.

基于自生成深度神经网络的4D航迹预测

Self-generated deep neural network based 4D trajectory prediction

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	刘世泽, 秦艳君, 王晨星, 苏琳, 柯其学, 罗海勇, 孙艺, 王宝会. 基于深度残差长短记忆网络交通流量预测算法[J]. 计算机应用, 2021, 41(6): 1566-1572.
[2]	陈凯, 于彦伟, 赵金东, 宋鹏. 基于城市交通监控大数据的工作位置推理方法[J]. 计算机应用, 2021, 41(1): 177-184.
[3]	龙洋洋, 陈玉玲, 辛阳, 豆慧. 基于联盟区块链的安全能源交易方案[J]. 计算机应用, 2020, 40(6): 1668-1673.
[4]	杜旭升, 于炯, 叶乐乐, 陈嘉颖. 基于图上随机游走的离群点检测算法[J]. 计算机应用, 2020, 40(5): 1322-1328.
[5]	徐周波, 杨健, 刘华东, 黄文文. 基于XGBoost与拓扑结构信息的蛋白质复合物识别算法[J]. 计算机应用, 2020, 40(5): 1510-1514.
[6]	马董, 陈红梅, 王丽珍, 肖清. 空间亚频繁co-location模式的主导特征挖掘[J]. 计算机应用, 2020, 40(2): 465-472.
[7]	陈曦, 梅广, 张金金, 许维胜. 融合知识图谱和协同过滤的学生成绩预测方法[J]. 计算机应用, 2020, 40(2): 595-601.
[8]	李莎莎, 梁冬阳, 余杰, 纪斌, 马俊, 谭郁松, 吴庆波. 基于师门关系的研究团队挖掘算法[J]. 计算机应用, 2020, 40(11): 3198-3202.
[9]	孙鹤立, 张优优, 杨洲, 何亮, 贾晓琳. 基于时间线段树的城市可达区域搜索[J]. 计算机应用, 2020, 40(10): 2936-2941.
[10]	李博, 张晓, 颜靖艺, 李可威, 李恒, 凌玉龙, 张勇. 基于值差度量和聚类优化的K最近邻算法在银行客户行为预测中的应用[J]. 计算机应用, 2019, 39(9): 2784-2788.
[11]	纪丽娜, 陈凯, 于彦伟, 宋鹏, 王淑莹, 王成锐. 基于城市交通大数据的车辆类别挖掘及应用分析[J]. 计算机应用, 2019, 39(5): 1343-1350.
[12]	叶志宇, 冯爱民, 高航. 基于深度LightGBM集成学习模型的谷歌商店顾客购买力预测[J]. 计算机应用, 2019, 39(12): 3434-3439.
[13]	袁钟, 冯山. 基于邻域值差异度量的离群点检测算法[J]. 计算机应用, 2018, 38(7): 1905-1909.
[14]	陈辉, 蒋圭峰, 姜桂圆, 武继刚. 基于海量公交轨迹数据挖掘的地图匹配算法[J]. 计算机应用, 2018, 38(7): 1923-1928.
[15]	陈柳, 冯山. 正负关联规则两级置信度阈值设置方法[J]. 计算机应用, 2018, 38(5): 1315-1319.