摘要
船舶航行的轨迹数据因时间序列稀疏性导致经验路径集中出现分形维度突变,使得路径规划过度关注局部而增加整体耗时。为此,提出基于历史数据挖掘的低延时船舶路径规划方法。分析船舶历史航行数据时间序列,计算路径平均运行速度,构建历史航行经验路径集;针对历史航行经验路径集路径耗散的分形维度,生成初始避碰路径,构建初始避碰路径的目标函数,综合考虑风浪和水流对速度的影响,平滑处理路径耗散分形维度,平衡局部与全局优化,实现船舶最佳低延时航行路径的求解。实验证明,所提方法能够实现船舶路径低延时规划,路径平滑度高,耗时少。
Due to the sparsity of time series,the trajectory data of ship navigation exhibits fractal dimension mutations in the empirical path set,leading to excessive focus on local areas in path planning and increasing overall time consumption.Therefore,a low latency ship path planning method based on historical data mining is proposed.Analyze the time series of historical navigation data of ships,calculate the average running speed of paths,and construct a set of historical navigation experience paths;Based on the fractal dimension of path dissipation in the historical navigation experience path set,generate initial collision avoidance paths,construct the objective function of the initial collision avoidance path,comprehensively consider the influence of wind,waves,and water flow on velocity,smooth the fractal dimension of path dissipation,balance local and global optimization,and achieve the solution of the optimal low delay navigation path for ships.Experimental results have shown that the proposed method can achieve low delay planning of ship paths,high path smoothness,and low time consumption.
作者
李宗锋
罗汉江
齐林
LI Zongfeng;LUO Hanjiang;QI Lin(Big Data Academy,Qingdao Huanghai University,Qingdao 266427,China;School of Computer Science and Engineering,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《舰船科学技术》
北大核心
2025年第16期163-167,共5页
Ship Science and Technology
基金
山东省自然科学基金面上项目(ZR2024MF024)。
关键词
历史航行数据
经验路径集挖掘
低延时规划
代价函数
historical navigation data
experience path set mining
low latency planning
cost function
作者简介
李宗锋(1981-),男,硕士,副教授,研究方向为信息管理、数据挖掘。
