考虑异常数据及船舶行为的在线AIS轨迹压缩算法  

An online AIS trajectory compression algorithm considering abnormal data and ship behaviors

作　　者：张俊峰 吴双 ZHANG Junfeng;WU Shuang(Institute for Big Data Research,Liaoning University of International Business and Economics,Dalian 116052,China)

机构地区：[1]辽宁对外经贸学院大数据研究院,辽宁大连116052

出　　处：《大连工业大学学报》2024年第6期462-468,共7页Journal of Dalian Polytechnic University

摘　　要：为了提高船舶AIS轨迹数据的在线压缩效率,提出了一种考虑AIS轨迹异常数据及船舶停留、航行等行为特征的在线轨迹数据压缩算法(ASN)。通过判断空间阈值和时间阈值,利用双通道对AIS轨迹数据进行分割和连接,实现在线清洗异常数据;创建临时停留窗口对船舶处于停留行为下的AIS轨迹数据进行分割;改进滑动窗口算法以保留更为细致的船舶航行行为数据。选取舟山水域实际船舶AIS轨迹数据进行试验分析,在压缩率相同的条件下,对比了Sliding Window算法、OPW算法、OPW-TR算法与SQUISH-E(λ)算法。结果表明,ASN算法在多个性能指标上取得了更好的效果。当压缩率为90%时,相比于目前效果较好的SQUISH-E(λ)算法,长度损失率降低了67.9%,轨迹相似度提高了61.6%,平均SED误差减少了35.5%,平均方向误差降低了65.2%,平均速度误差减少了32.0%,有效提升了AIS轨迹数据的在线压缩效率。To improve the online compression efficiency of ship AIS trajectory data,an online trajectory data compression algorithm(ASN algorithm)was proposed,which considered the abnormal data of AIS trajectory and the behavior characteristics of ship stay and navigation.By judging the spatial threshold and time threshold,the AIS trajectory data was segmented and connected by dual channels to realize online cleaning of abnormal data.A temporary stay window was created to segment the AIS trajectory data of the ship under the stay behavior.The sliding window algorithm was improved to retain more detailed data of ship navigation behavior.The AIS trajectory data of actual ships in Zhoushan waters were selected for experimental analysis.Under the condition of the same compression rate,Sliding Window algorithm,OPW algorithm,OPW-TR algorithm and SQUISH-E(λ)algorithm were compared.The results showed that the ASN algorithm achieved better results in many performance indexes.When the compression ratio was 90%,the length loss rate was reduced by 67.9%,the trajectory similarity was increased by 61.6%,the average SED error was reduced by 35.5%,the average direction error was reduced by 65.2%,and the average speed error was reduced by 32.0%,compared with the current SQUISH-E(λ)algorithm with better performance.The online compression efficiency of AIS trajectory data is effectively improved.

关 键 词：AIS数据 船舶行为特征 滑动窗口 轨迹压缩 

分 类 号：U644.1[交通运输工程—船舶及航道工程]