基于强化学习的自动帆船路径规划算法研究  

Research on Path Planning of Automatic Sailboat Based on Reinforcement Learning

作　　者：荆纬 李浩 咸琳涛 刘源 郭忠文[1] Jing Wei;Li Hao;Xian Lintao;Liu Yuan;Guo Zhongwen(College of Information and Engineering, Ocean University of China, Qingdao 266100, China)

机构地区：[1]中国海洋大学信息科学与工程学院,山东青岛266100

出　　处：《中国海洋大学学报（自然科学版）》2021年第8期79-87,共9页Periodical of Ocean University of China

基　　金：国家自然科学基金项目(61827810);山东省自然科学基金项目(ZR2018BF006)资助。

摘　　要：作为自动水面航行器的重要分支之一,自动航行帆船在执行长期海事任务时具有低能耗的优势,但其航行过程受到环境因素的影响很大。针对以上情况,本文考虑了自动帆船的自身运动模型,以及在航行时受到的海风、海流和障碍物的影响,提出了自动帆船从起始点至目标点的路径规划算法。该算法通过帆船的平面运动模型来计算环境因素的影响,再通过强化学习中的Q-learning算法实现对于海上两点间的路径规划并同时实现规避障碍物。通过仿真实验证明了本文提出的自动航行帆船的路径规划算法是可行的。As one of the important branches of the automatic surface vehicle,the automatic sailboat has the advantage of low energy consumption when it carries out the long-run marine missions.Meanwhile,it’s greatly affected by environmental factors,such as sea winds,sea currents,obstaclesand the motion model of automatic sailboat on the navigation.Considering these circumstances,this paper studiesthe external influencing factors and proposes a path planning algorithm.The proposed algorithm calculates the influence of environmental factors through the planar motion model of sailboat.Thenit realizes the path planning and the obstacle avoidance among two given points at sea by Q-learning algorithm of reinforcement learning.The simulation results show that the proposed algorithm is feasible.

关 键 词：帆船 自动航行 路径规划 强化学习 海流 海风 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]