基于AIS数据的船舶冲突概率检测方法
A Method for Estimating Ship Motion Conflict Risk Based on AIS Information
-
摘要: 针对高交通量的港口水域和动态航行场景下有效的船舶碰撞风险感知问题,提出一种冲突概率检测方法。基于船舶自动识别系统(Automatic Identification System, AIS)数据挖掘(Data Mining, DM)提取船长、航速和相对方位这3者与最小安全距离(Safe Distance of Approach, SDAmin)的关系,构建动态领域模型;在此基础上,考虑港口水域船舶航行的动态性和船舶位置预测的不确定性,基于蒙特卡洛仿真根据不确定性分量的概率密度函数,在船舶不确定性运动方程模型中插入每个随机样本点,利用确定性求解器对冲突概率进行求解。结果表明:该方法能在高交通量的复杂动态环境下及时、准确地检测碰撞风险,可为船舶冲突的判断和智能避碰提供理论依据和指导。Abstract: The method for estimating ship motion conflict is developed for ship collision risks perception in crowded harbor waters or dynamic navigation scenarios. AIS Data are mined to find information about involved ships, such as the relationship between 3 variables(the captain, the speed and relative azimuth) and SDAmin. The dynamic domain model is built based on the information. The uncertainty of ship navigation dynamics and the ship position predictions is taken into account by incorporating random samples in the ship motion model according to the probability density of the uncertainty components, which embodies the Monte Carlo simulation. This allows the conflict risk problem to be solved with deterministic algorithm. Tests show that the method can give accurate risk estimates in a timely manner.