摘要
Abstract
Ocean currents,which have a substantial impact on the navigation of low-speed,underdriven AUVs,can in-crease navigation time,raise energy consumption,and change the navigation trajectory.Therefore,planning an optimal navigation route that accounts for the disturbance of ocean currents is of considerable importance.This study mainly analyzes the mechanism by which ocean currents influence AUVs and proposes an improved DQN path planning al-gorithm based on the prioritized experience replay method.This modification addresses the problem of overestimation,which is a common issue when using a traditional DQN path planning algorithm.Additionally,the action design and re-ward functions are optimized.Path planning simulations are conducted in a 3D ocean environment,which is established based on S57 chart data and ocean current data provided by Earth&Space Research.Experimental results show that the improved DQN algorithm generates a more effective global path planning,offering a navigation route that minimizes time and energy consumption.This work provides a valuable reference for underwater AUV navigation,fully consider-ing the impact of ocean current disturbances.关键词
自主水下运载器/强化学习/洋流干扰/路径规划/三维海洋环境/强化Q网络/S57海图/奖励函数Key words
automatic underwater vehicle/reinforcement learning/ocean current disturbance/path planning/3D marine environment/deep Q-network/S57 charts/reward function分类
计算机与自动化
