摘要
针对船舶会遇过程中的避让行为差异性问题,本文提出一种基于稀疏反演理论的船舶行为不确定性分析方法。通过碰撞事故调查报告中船舶自动识别系统数据的情景反演,提出船舶会遇过程中包括避让动机、时机、方式、幅度和效果等行为表征;针对不同会遇局面的行为程度特点,构建神经网络辨识船舶避让行为,利用信息熵测量不同会遇阶段下行为的不确定性程度;结合多起碰撞事故调查报告的样本数据,揭示碰撞事故集下船舶避让行为的不确定性特征。结果表明:船舶具有不同局面不同避让责任的差异化避让行为特征,且随两船相互驶近下避让行为的方式、幅度和效果等不确定性逐渐增大。基于稀疏反演理论的数据分析方法可量化船舶避让行为的时空分异特性,为船舶会遇避让的决策方案提供智能化分析的思路。
To address the variability in ships′avoidance behaviors during encounters,a sparse inversion-based method for analyzing the uncertainty of ship avoidance behavior(SAB)was proposed.First,a method for identifying SAB during ship encounters was introduced by inverting AIS data from collision accident reports.Second,a neural network was constructed based on the behavioral characteristics of different situations to identify SAB,and information entropy was used to measure the uncertainty of SAB at different stages.Finally,by analyzing multiple accident reports,uncertainties were revealed in several aspects of SAB,including ship avoidance motivation,opportunity,pattern,extent,and effectiveness.The results indicate that ships display differentiated avoidance behavior characteristics depending on the situation and avoidance responsibilities,revealing a gradually increasing uncertainty in the pattern,extent,and effectiveness of SAB as two ships approach each other.The data analysis method based on sparse inversion can quantify the spatiotemporal differentiation of SAB,providing a new approach for the intelligent analysis of decision-making plans within ship collision avoidance frameworks.
作者
胡甚平
陈炎
朱清华
韩冰
HU Shenping;CHEN Yan;ZHU Qinghua;HAN Bing(Merchant Marine College,Shanghai Maritime University,Shanghai 201306,China;Shanghai Ship and Shipping Research Institute Co.,Ltd.,Shanghai 200135,China)
出处
《哈尔滨工程大学学报》
北大核心
2025年第9期1709-1718,共10页
Journal of Harbin Engineering University
基金
国家自然科学基金项目(52272353).
关键词
水上交通
碰撞事故
船舶避让行为
不确定性
稀疏反演理论
数据驱动
神经网络
信息熵
marine traffic
collision accident
ships′avoidance behavior
uncertainty analysis
sparse inversion theory
data-driven
neural network
information entropy
