复杂开放水域受气象、海况、交通流量等因素影响，其他船舶航行轨迹不确定，增加了航迹规划难度。为此，提出船舶避障航迹自适应规划方法。将其他船舶视为动态障碍物，用状态转移方程和观测方程自适应跟踪，借助强化学习构建环境模型，把船舶位置、动态与静态障碍物位置映射到二维坐标系，真实反映水域情况。将船舶动作空间离散为 8 个方向，泛化激励函数得到自适应非线性分段函数。船舶可据此根据自身及环境变化自适应调整航迹规划策略。实验表明，该方法构建的船舶运动状态模型获取的航向角与 GPS 数值高度吻合，动态障碍物跟踪能力强，规划的航迹能精准避障，应用性佳。
Complex open waters are affected by factors such as weather, sea conditions, and traffic flow, and the navigation trajectories of other ships are uncertain, which increases the difficulty of trajectory planning. Therefore, an adaptive planning method for ship obstacle avoidance trajectory is proposed. Treat other ships as dynamic obstacles, adaptively track them using state transition equations and observation equations, and use reinforcement learning to construct an environmental model. Map the ship's position, dynamic and static obstacle positions to a two-dimensional coordinate system to truly reflect the water situation. Discretize the ship’s motion space into 8 directions and generalize the excitation function to obtain an adaptive nonlinear piecewise function. Ships can adaptively adjust their trajectory planning strategies based on their own and environmental changes. The experiment shows that the heading angle obtained by the ship motion state model constructed by this method is highly consistent with GPS values, has strong dynamic obstacle tracking ability, and the planned trajectory can accurately avoid obstacles, with good applicability.
2025,47(21): 27-31 收稿日期：2025-5-28
DOI：10.3404/j.issn.1672-7649.2025.21.005
分类号：U675.7
基金项目：中国交通教育研究会2022–2024年度教育研究课题（JT2022YB144）
作者简介：陈留远(1983-),男,无限航区一等船长,研究方向为航海技术
参考文献：
[1] 颜明重, 李琛, 朱大奇. 基于生物启发神经网络的无人水面艇实时避障路径规划[J]. 上海海事大学学报, 2024, 45(3): 10-15+48.
YAN M Z, LI C, ZHU D Q. Real-time obstacle avoidance path planning based on bio-inspired neural network for unmanned surface vehicles[J]. Journal of Shanghai Maritime University, 2024, 45(3): 10-15+48.
[2] 李光泽, 俞万能, 廖卫强, 等. 基于动态窗口法的近海水域船舶避障算法研究[J]. 中国舰船研究, 2024, 19(1): 230-237.
LI G Z, YU W N, LIAO W Q, et al. Obstacle avoidance algorithm for ships in complex waters based on dynamic window approach[J]. Chinese Journal of Ship Research, 2024, 19(1): 230-237.
[3] 谭智坤, 张隆辉, 刘正锋, 等. 融合改进动态窗口法与速度障碍法的无人船局部路径规划[J]. 船舶力学, 2023, 27(3): 311-322.
TAN Z K, ZHANG L H, LIU Z F, et al. Local path planning for USVs based on the fusion algorithm of improved dynamic window approach and velocity obstacle algorithm[J]. Journal of Ship Mechanics, 2023, 27(3): 311-322.
[4] 周寅飞, 张立华, 贾帅东, 等. 最大可航窗口序列约束贝塞尔曲线的无人船自主航行航线规划方法[J]. 武汉大学学报(信息科学版), 2024, 49(7): 1224-1236.
ZHOU Y F, ZHANG L H, JIA S D, et al. Autonomous navigation route planning method of unmanned ship based on bessel curves constrained by maximum navigable window sequence[J]. Geomatics and Information Science of Wuhan University, 2024, 49(7): 1224-1236.
[5] 刘文博, 鲁阳, 薛彦卓. 面向冰区船舶航线的混合规划算法[J]. 哈尔滨工程大学学报, 2024, 45(11): 2065-2074.
LIU W B, LU Y, XUE Y Z. A hybrid route planning algorithm for ship routing in ice areas[J]. Journal of Harbin Engineering University, 2024, 45(11): 2065-2074.
[6] 庹玉龙, 康彩霞, 王莎莎, 等. 未知干扰下多船相邻交叉耦合同步编队避障控制[J]. 交通运输工程学报, 2023, 23(6): 314-326.
TUO Y L, KANG C X, WANG S S, et al. Adjacent cross-coupling synchronous formation control with collision avoidance for multiple ships under unknown disturbances[J]. Journal of Traffic and Transportation Engineering, 2023, 23(6): 314-326.
[7] 盛金亮, 董早鹏, 匡文琪, 等. 基于改进DWA的无人艇编队避障及MPC轨迹跟踪[J]. 大连海事大学学报, 2024, 50(4): 12-21.
SHENG J L, DONG Z P, KUANG W Q, et al. Obstacle avoidance and MPC trajectory tracking of unmanned surface vehicle formation based on improved DWA[J]. Journal of Dalian Maritime University, 2024, 50(4): 12-21.
[8] 唐裕峰, 宋瑶, 武浩正, 等. 雷达与AIS对海上低速目标航迹关联的特征选择[J]. 计算机仿真, 2023, 40(12): 193-199.
TANG Y F, SONG Y, WU H Z, et al. Feature selection for correlation of radar and ais with low speed target tracks at sea[J]. Computer Simulation, 2023, 40(12): 193-199.
