面向智能航行的货船“航行脑”概念设计

摘要

提出“航行脑”系统的概念设计.该系统是服务于船舶智能航行的人工智能系统,由感知、认知和决策执行等3个功能空间组成.“感知空间”获取船舶在航环境和自身状态信息;“认知空间”根据感知的信息抽象出航行态势,实现自身状态辨识,最终基于人工驾驶记录和机器学习建立智能船舶驾驶行为谱;“决策执行空间”利用“感知空间”反馈的信息修正“认知空间”的态势认知,在驾驶行为谱的支持下实现对智能船舶的鲁棒控制.分析“航行脑”系统的“感知空间”“认知空间”和“决策执行空间”等关键技术,并展望“航行脑”系统在智能货船上的应用,以实现货船的智能航行,达到减少配员、降低排放和提高船舶航行安全性的目的.%The concept design of Navigation Brain System (NBS) for intelligent ship is put forward.The NBS is composed of three subsystems,a percepion module,a cognizing module,and a decision and manipulation module.The perception module is designed for obtaining the environmental information and the status of cargo ship itself.The cognizing module identifies the navigational situation from the perception information and builds up the behavior patterns of the ship through machine learning.The decision and manipulation module controls the ship and updates the output of the cognizing module according to the feedback of the perception module.The key technologies of a navigation brain system are presented.