The invention discloses an intelligent ship navigation dense area collision avoidance decision-making method based on target evaluation learning. The method comprises the steps that conventional tuples in the Markov decision-making process are expanded to construct a collision avoidance decision-making model used for an intelligent ship in a navigation dense area under target guidance; and based on a post experience playback idea, establishing a target evaluation learning algorithm of the collision avoidance decision of the intelligent ship in the navigation dense area to train the collision avoidance decision model, so that the intelligent ship can perform decision output in the navigation dense area. According to the method, the decision-making process of the intelligent ship in the densely navigable area is converted into a target-oriented learning model, so that more reliable and intelligent technical support is provided for safe navigation of the ship in a complex marine environment.
本发明公开了一种基于目标评价学习的智能船通航密集区避碰决策方法,包括:对马尔可夫决策过程中的常规元组进行扩充构建用于目标导向下智能船在通航密集区的避碰决策模型;基于事后经验回放思想,建立智能船在通航密集区下避碰决策的目标评价学习算法对避碰决策模型进行训练,使得智能船舶能够在通航密集区域进行决策输出。本发明通过将智能船在通航密集区域的决策过程转化为目标导向的学习模型,为船舶在复杂海上环境中的安全航行提供了更为可靠、智能的技术支持。
Intelligent ship navigation dense area collision avoidance decision-making method based on target evaluation learning
一种基于目标评价学习的智能船通航密集区避碰决策方法
2024-07-12
Patent
Electronic Resource
Chinese
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS |
Ship intelligent navigation collision avoidance efficiency verification and evaluation method
European Patent Office | 2021
|Intelligent collision avoidance method and system for ship navigation
European Patent Office | 2024
|European Patent Office | 2023
|Research on Collision Avoidance Decision-Making of Unmanned Ship Navigation in Fairway
Springer Verlag | 2023
|