基于上下文自编码的船舶行为语义表征

马杰; 何沐蓉; 贾承丰; 李文楷; 张煜

doi:10.19818/j.cnki.1671-1637.2022.04.026

基于上下文自编码的船舶行为语义表征

doi: 10.19818/j.cnki.1671-1637.2022.04.026

马杰^{1, 2, 3,},
何沐蓉¹,
贾承丰¹,
李文楷¹,
张煜^{2, 3, 4}

1.
武汉理工大学航运学院，湖北武汉 430063
2.
武汉理工大学内河航运技术湖北省重点实验室，湖北武汉 430063
3.
武汉理工大学国家水运安全工程技术研究中心，湖北武汉 430063
4.
武汉理工大学交通与物流工程学院，湖北武汉 430063

基金项目:

国家重点研发计划 2021YFB3901504

国家自然科学基金项目 52271366

国家自然科学基金项目 51679182

详细信息

作者简介:
马杰(1978-)，男，湖北武汉人，武汉理工大学教授，工学博士，从事信息融合感知与船舶智能导航研究

中图分类号: U675
计量
- 文章访问数: 880
- HTML全文浏览量: 315
- PDF下载量: 114
- 被引次数: 0
出版历程
- 收稿日期: 2022-02-09
- 网络出版日期: 2022-10-08
- 刊出日期: 2022-08-25

Semantic representation of ship behavior based on context autoencoder

MA Jie^{1, 2, 3
,},
HE Mu-rong¹,
JIA Cheng-feng¹,
LI Wen-kai¹,
ZHANG Yu^{2, 3, 4}

1.
School of Navigation, Wuhan University of Technology, Wuhan 430063, Hubei, China
2.
Hubei Key Laboratory of Inland Shipping Technology, Wuhan University of Technology, Wuhan 430063, Hubei, China
3.
National Engineering Research Center for Water Transportation Safety, Wuhan University of Technology, Wuhan 430063, Hubei, China
4.
School of Transportation and Logistics Engineering, Wuhan University of Technology, Wuhan 430063, Hubei, China

Funds:

National Key Research and Development Program of China 2021YFB3901504

National Natural Science Foundation of China 52271366

National Natural Science Foundation of China 51679182

More Information

Author Bio:
MA Jie(1978-), male, professor, PhD, majie@whut.edu.cn

摘要

摘要: 考虑船舶行为的时序相关性，提出了一种基于上下文自编码的船舶行为语义表征(SRCAE)模型；提取船舶经度、纬度、航速、航向等行为特征参量，建立了行为特征序列；借助连续词袋模型将行为特征序列划分为中心船舶行为和上下文船舶行为，利用深度自编码网络构建了船舶上下文行为的语义表征模型，将得到的中心船舶行为编码作为表征向量输出，通过聚类算法构建船舶行为词典；选取长江口南槽交汇水域作为研究对象，利用船舶自动识别系统产生的数据对提出的模型和方法进行了验证。分析结果表明：所提出的SRCAE模型能有效表征船舶行为之间的上下文联系，与传统自编码器和长短期记忆网络自编码器等模型相比SRCAE模型具有更低的表征误差；分别采用k均值(k-Means)、高斯混合模型(GMM)与核k均值(Kernel k-Means)3种聚类算法提取船舶行为词典，与原始数据相比SRCAE模型产生的表征向量更易于区分不同船舶行为模式，其中k-Means效果最优，轮廓系数、卡林斯基-哈拉巴斯指数和戴维森堡丁指数指标分别达到了0.384、18.308、0.531，共产生转向加速、转向减速、直行加速、直行减速等30种复合行为，有效提取了不同行为模式下船舶行为词组合关系。
- 船舶自动识别系统 /
- 船舶行为 /
- 自编码器 /
- 连续词袋模型 /
- 语义表征
Abstract: Considering the temporal correlation of ship behavior, a semantic representation model based on the context autoencoder (SRCAE) was proposed for ship behavior. The behavioral feature parameters, such as the longitude, latitude, speed, as well as the course, were extracted to establish the behavioral feature sequence. The behavioral feature sequence was divided into the center ship behavior and context ship behavior via the continuous bag-of-words (CBOW) model. The deep autoencoder (AE) networks were utilized to construct the semantic representation model of context ship behavior, and the encoded center ship behavior obtained from the model was output as the representation vector. The clustering algorithm was employed to establish the ship behavior dictionary. The South Passage Intersection Water of the Yangtze Estuary was selected as the research object, and the data from the automatic identification system (AIS) for ships were employed for verification of the proposed model and method. Analysis results show that the context relationships between ship behaviors can be effectively represented by the proposed SRCAE model, and the representation error of the SRCAE model is lower than that of the traditional AE model and long short-term memory autoencoder (LSTMAE) model. Three clustering algorithms, namely, k-means, Gaussian mixture model (GMM), and kernel k-means, were used to extract the ship behavior dictionary. Compared with the original data, the representation vectors generated by the SRCAE model are easier to distinguish different ship behavior patterns, among which the effect of k-means is the best, and the Silhouette coefficient (SC), Calinski-Harabasz index (CHI), and Davies-Bouldin index (DBI) of k-means reach 0.384, 18.308, and 0.531, respectively. A total of 30 types of composite behaviors are generated, such as steering acceleration, steering deceleration, straight-ahead acceleration, straight-ahead deceleration, and so on and the combination relationships of ship behavior words under different behavior patterns are effectively extracted.
- ship automatic identification system /
- ship behavior /
- autoencoder /
- continuous bag-of-words model /
- semantic representation

HTML全文

图 1 SOG语义转换

Figure 1. Semantic transformation of SOG

下载: 全尺寸图片幻灯片

图 2 SRCAE模型框架

Figure 2. Framework of SRCAE model

下载: 全尺寸图片幻灯片

图 3 中心船舶行为和上下文船舶行为提取

Figure 3. Extraction of center ship behavior and context ship behavior

下载: 全尺寸图片幻灯片

图 4 重构全局行为的行为特征参量误差(244010871)

Figure 4. Parameter errors of behavior characteristics for reconstructing global behavior (244010871)

下载: 全尺寸图片幻灯片

图 5 重构全局行为的行为特征参量误差(255805790)

Figure 5. Parameter errors of behavior characteristics for reconstructing global behavior (255805790)

下载: 全尺寸图片幻灯片

图 6 重构全局行为的行为特征参量误差(255805792)

Figure 6. Parameter errors of behavior characteristics for reconstructing global behavior (255805792)

下载: 全尺寸图片幻灯片

图 7 重构全局行为的行为特征参量误差(351453000)

Figure 7. Parameter errors of behavior characteristics for reconstructing global behavior (351453000)

下载: 全尺寸图片幻灯片

图 8 船舶行为表征算法的性能对比

Figure 8. Performance comparison of ship behavior representation algorithms

下载: 全尺寸图片幻灯片

图 9 三种聚类算法在原始轨迹数据与表征向量中聚类结果

Figure 9. Clustering results of three clustering algorithms on raw trajectory data and representation vector

下载: 全尺寸图片幻灯片

图 10 30类聚类中心在经纬度中的分布

Figure 10. Distribution of 30 cluster centers in latitude and longitude

下载: 全尺寸图片幻灯片

图 11 船舶行为场景实例

Figure 11. Examples of ship behavior

下载: 全尺寸图片幻灯片

表 1 三种聚类算法在表征向量中聚类效果对比

Table 1. Comparison of three clustering algorithms in representation vector

方法	指标
方法	SC	CHI	DBI
k-Means	0.384	18.308	0.531
GMM	0.368	18.665	1.675
核k-Means	0.376	17.874	0.792

下载: 导出CSV

表 2 6类船舶行为词典统计描述

Table 2. Statistical descriptions of six kinds of ship behavior dictionary

类别	行为参数	均值	标准差	上分位数	中位数	下分位数
B₂	经度/(°)	122.561	0.022	122.545	122.549	122.581
	纬度/(°)	31.017	0.022	30.992	31.028	31.038
	对地航速/kn	10.429	2.462	8.300	10.600	12.200
	对地航向/(°)	271.868	115.087	278.730	287.880	357.760
B₃	经度/(°)	122.568	0.024	122.546	122.551	122.592
	纬度/(°)	31.011	0.027	30.989	31.018	31.038
	对地航速/kn	10.173	2.439	8.300	10.100	11.400
	对地航向/(°)	187.562	144.453	2.920	275.320	287.000
B₅	经度/(°)	122.531	0.032	122.543	122.545	122.548
	纬度/(°)	30.949	0.030	30.930	30.935	30.948
	对地航速/kn	7.664	3.477	6.400	7.300	8.270
	对地航向/(°)	292.245	120.646	282.490	356.000	358.180
B₁₄	经度/(°)	122.543	0.008	122.543	122.545	122.548
	纬度/(°)	30.968	0.012	30.958	30.966	30.974
	对地航速/kn	9.049	3.865	7.600	8.820	10.000
	对地航向/(°)	318.687	100.557	353.790	357.000	358.280
B₁₉	经度/(°)	122.524	0.017	122.512	122.520	122.543
	纬度/(°)	30.969	0.016	30.958	30.968	30.986
	对地航速/kn	9.173	3.626	7.950	8.840	9.800
	对地航向/(°)	145.350	97.907	103.840	172.680	180.620
B₂₈	经度/(°)	122.472	0.015	122.462	122.469	122.479
	纬度/(°)	31.004	0.014	30.997	31.010	31.013
	对地航速/kn	10.387	2.686	8.500	9.700	12.800
	对地航向/(°)	231.500	78.093	115.530	279.830	282.250

下载: 导出CSV

参考文献(31)

[1]	朱姣, 刘敬贤, 陈笑, 等. 基于轨迹的内河船舶行为模式挖掘[J]. 交通信息与安全, 2017, 35(3): 107-116, 132. doi: 10.3963/j.issn.1674-4861.2017.03.014 ZHU Jiao, LIU Jing-xian, CHEN Xiao, et al. Behavior pattern mining of inland vessels based on trajectories[J]. Journal of Transport Information and Safety, 2017, 35(3): 107-116, 132. (in Chinese) doi: 10.3963/j.issn.1674-4861.2017.03.014
[2]	文元桥, 张义萌, 黄亮, 等. 基于语义的船舶行为动态推理机制[J]. 中国航海, 2019, 42(3): 34-39, 50. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGHH201903008.htm WEN Yuan-qiao, ZHANG Yi-meng, HUANG Liang, et al. Mechanism of ship behavior dynamic reasoning based on semantics[J]. Navigation of China, 2019, 42(3): 34-39, 50. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGHH201903008.htm
[3]	TU En-mei, ZHANG Guang-hao, RACHMAWATI L, et al. Exploiting AIS data for intelligent maritime navigation: a comprehensive survey from data to methodology[J]. IEEE Transactions on Intelligent Transportation Systems, 2017, 19(5): 1559-1582.
[4]	LEI D, FLORIS G, PENTTI K. Review and analysis of methods for assessing maritime waterway risk based on non-accident critical events detected from AIS data[J]. Reliability Engineering and System Safety, 2020, 200: 106933. doi: 10.1016/j.ress.2020.106933
[5]	YANG Dong, WU Ling-xiao, WANG Shuai-an, et al. How big data enriches maritime research—a critical review of automatic identification system (AIS) data applications[J]. Transport Reviews, 2019, 39(6): 755-773. doi: 10.1080/01441647.2019.1649315
[6]	甄荣, 邵哲平, 潘家财. 基于AIS数据的船舶行为特征挖掘与预测: 研究进展与展望[J]. 地球信息科学学报, 2021, 23(12): 2111-2127. doi: 10.12082/dqxxkx.2021.210495 ZHEN Rong, SHAO Zhe-ping, PAN Jia-cai. Advance in character mining and prediction of ship behavior based on AIS data[J]. Journal of Geo-Information Science, 2021, 23(12): 2111-2127. (in Chinese) doi: 10.12082/dqxxkx.2021.210495
[7]	陶阳, 毛喆, 盛萍, 等. 基于船舶行为的武汉长江大桥水域货船通航规律研究[J]. 交通信息与安全, 2018, 36(1): 49-56. doi: 10.3963/j.issn.1674-4861.2018.01.007 TAO Yang, MAO Zhe, SHENG Ping, et al. A study of regularity of navigation patterns of cargo ships at the waterways near Wuhan Yangtze River Bridge based on ship maneuvering behavior[J]. Journal of Transport Information and Safety, 2018, 36(1): 49-56. (in Chinese) doi: 10.3963/j.issn.1674-4861.2018.01.007
[8]	WU Xing, MEHTA A L, ZALOOM V A, et al. Analysis of waterway transportation in Southeast Texas waterway based on AIS data[J]. Ocean Engineering, 2016, 121: 196-209. doi: 10.1016/j.oceaneng.2016.05.012
[9]	XIAO Fang-liang, LIGTERINGEN H, VAN-GULIJK C, et al. Comparison study on AIS data of ship traffic behavior[J]. Ocean Engineering, 2015, 95: 84-93. doi: 10.1016/j.oceaneng.2014.11.020
[10]	甄荣, 邵哲平, 潘家财, 等. 基于AIS信息的航道内船舶速度分布统计分析[J]. 集美大学学报(自然科学版), 2014, 19(4): 274-278. doi: 10.3969/j.issn.1007-7405.2014.04.006 ZHEN Rong, SHAO Zhe-ping, PAN Jia-cai, et al. Statistical analysis of distribution of ship speed within the fairway based on AIS data[J]. Journal of Jimei University (Natural Science), 2014, 19(4): 274-278. (in Chinese) doi: 10.3969/j.issn.1007-7405.2014.04.006
[11]	HUANG Liang, WEN Yuan-qiao, GUO Wei, et al. Mobility pattern analysis of ship trajectories based on semantic transformation and topic model[J]. Ocean Engineering, 2020, 201: 107092. doi: 10.1016/j.oceaneng.2020.107092
[12]	朱飞祥, 张英俊, 高宗江. 基于数据挖掘的船舶行为研究[J]. 中国航海, 2012, 35(2): 50-54. doi: 10.3969/j.issn.1000-4653.2012.02.011 ZHU Fei-xiang, ZHANG Ying-jun, GAO Zong-jiang. Research on ship behaviors based on data mining[J]. Navigation of China, 2012, 35(2): 50-54. (in Chinese) doi: 10.3969/j.issn.1000-4653.2012.02.011
[13]	ZHANG Wei-bin, GOERLANDT F, KUJALA P, et al. An advanced method for detecting possible near miss ship collisions from AIS data[J]. Ocean Engineering, 2016, 124: 141-156. doi: 10.1016/j.oceaneng.2016.07.059
[14]	BENGIO Y, COURVILLE A, VINCENT P. Representation learning: a review and new perspectives[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(8): 1798-1828. doi: 10.1109/TPAMI.2013.50
[15]	WANG Sheng, BAO Zhi-feng, CULPEPPER J S, et al. A survey on trajectory data management, analytics, and learning[J]. ACM Computing Surveys (CSUR), 2021, 53(2): 1-36.
[16]	LI Li, LI Xin, YANG Yuan, et al. Indoor tracking trajectory data similarity analysis with a deep convolutional autoencoder[J]. Sustainable Cities and Society, 2019, 45: 588-595. doi: 10.1016/j.scs.2018.12.025
[17]	WANG Wei, XIA Feng, NIE Han-song, et al. Vehicle trajectory clustering based on dynamic representation learning of internet of vehicles[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 22(6): 3567-3576. doi: 10.1109/TITS.2020.2995856
[18]	ZHOU Yang, DAAMEN W, VELLINGA T, et al. Ship classification based on ship behavior clustering from AIS data[J]. Ocean Engineering, 2019, 175(1): 176-187.
[19]	YAO Di, ZHANG Chao, ZHU Zhi-hua, et al. Trajectory clustering via deep representation learning[C]//IEEE. 2017 International Joint Conference on Neural Networks (IJCNN). New York: IEEE, 2017: 3880-3887.
[20]	ZHANG Rui, XIE Peng, JIANG Huang-bo, et al. Clustering noisy trajectories via robust deep attention auto-encoders[C]//IEEE. 2019 20th IEEE International Conference on Mobile Data Management (MDM). New York: IEEE, 2019: 63-71.
[21]	谢鹏. 基于深度学习的轨迹数据中移动模式发现方法研究[D]. 武汉: 武汉理工大学, 2019. XIE Peng. Research on moving pattern discovery method in trajectory data based on deep learning[D]. Wuhan: Wuhan University of Technology, 2019. (in Chinese)
[22]	常吉亮, 谢磊, 赵建伟, 等. 基于VAE-LSTM模型的航迹异常检测算法[J]. 交通信息与安全, 2020, 38(6): 1-8. https://www.cnki.com.cn/Article/CJFDTOTAL-JTJS202006002.htm CHANG Ji-liang, XIE Lei, ZHAO Jian-wei, et al. An anomaly detection algorithm for ship trajectory data based on VAE-LSTM model[J]. Journal of Transport Information and Safety, 2020, 38(6): 1-8. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JTJS202006002.htm
[23]	WEN Yuan-qiao, ZHANG Yi-meng, HUANG Liang, et al. Semantic modelling of ship behavior in harbor based on ontology and dynamic Bayesian network[J]. ISPRS International Journal of Geo-Information, 2019, DOI: 10.3390/ijgi8030107.
[24]	PARENT C, SPACCAPIETRA S, RENSO C, et al. Semantic trajectories modeling and analysis[J]. ACM Computing Surveys (CSUR), 2013, 45(4): 1-32.
[25]	MIKOLOV T, CHEN K, CORRADO G, et al. Efficient estimation of word representations in vector space[J]. arXiv, 2013, Bibcode: 2013arXiv1301.3781M.
[26]	甄荣, 金永兴, 胡勤友, 等. 基于AIS信息和BP神经网络的船舶航行行为预测[J]. 中国航海, 2017, 40(2): 6-10. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGHH201702002.htm ZHEN Rong, JIN Yong-xing, HU Qin-you, et al. Vessel behavior prediction based on AIS data and BP neural network[J]. Navigation of China, 2017, 40(2): 6-10. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZGHH201702002.htm
[27]	刘娇, 史国友, 杨学钱, 等. 基于DE-SVM的船舶航迹预测模型[J]. 上海海事大学学报, 2020, 41(1): 34-39, 115. https://www.cnki.com.cn/Article/CJFDTOTAL-SHHY202001006.htm LIU Jiao, SHI Guo-you, YANG Xue-qian, et al. Ship trajectory prediction model based on DE-SVM[J]. Journal of Shanghai Maritime University, 2020, 41(1): 34-39, 115. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-SHHY202001006.htm
[28]	FANG Zhi-xiang, YU Hong-chu, KE Ran-xuan, et al. Automatic identification system-based approach for assessing the near-miss collision risk dynamics of ships in ports[J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 20(2): 534-543.
[29]	LIN J, KEOGH E, WEI L, et al. Experiencing SAX: a novel symbolic representation of time series[J]. Data Mining and Knowledge Discovery, 2007, 15(2): 107-144.
[30]	KRASNOV F, SEN A. The number of topics optimization: clustering approach[J]. Machine Learning and Knowledge Extraction, 2019, 1(1): 416-426.
[31]	WANG Wen-shuo, RAMESH A, ZHU Jia-cheng, et al. Clustering of driving encounter scenarios using connected vehicle trajectories[J]. IEEE Transactions on Intelligent Vehicles, 2020, 5(3): 485-496.