基于动态多实体贝叶斯网络的无人艇任务决策研究  

Mission decision-making of USV based on dynamic multi-entity Bayesian network

作　　者：刘力源 周洪涛[1] 苏厚胜[1] LIU Li-yuan;ZHOU Hong-tao;SU Hou-sheng(School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]华中科技大学人工智能与自动化学院,湖北武汉430074

出　　处：《舰船科学技术》2023年第8期78-83,共6页Ship Science and Technology

基　　金：国家自然科学基金资助项目(61991412)。

摘　　要：任务决策是实现无人艇智能自主的关键环节,其存在着场景多样性、状态不确定性、约束动态性等问题。为此,本文提出动态多实体贝叶斯网络模型。首先设计了面向无人艇的语义推理框架,对无人艇本体信息进行知识表示。进一步,采用概率本体语言方法描述不确定性知识,同时扩展动态贝叶斯网络的结构,从而提高了无人艇任务决策在不确定性因素和时序因素影响下的推理能力。最后针对岛礁区域防守的任务场景,与多实体贝叶斯网络模型进行对比,结果表明DMEBN模型在动态条件下具有执行策略的连续性,验证了模型的可行性和有效性。Mission decision-making is the key component to realize the intelligent autonomy of unmanned surface vehicles,which has the problems of scenario diversity,state uncertainty,and constraint dynamics.To solve the above limitations,the dynamic multi-entity Bayesian network(DMEBN)model is proposed.In order to perform knowledge representation of unmanned surface vehicle ontology information,a semantic reasoning framework for unmanned surface vehicles is designed.The probabilistic ontology language is used to describe uncertainty knowledge while extending the structure of dynamic Bayesian networks,thus enhancing the reasoning ability of unmanned surface vehicle decision-making under the influence of uncertainty factors and temporal sequence.The model is applied to the mission scenario of island area defense and compared with the model of multi-entity Bayesian network to show the continuity of the DMEBN model in executing strategies under dynamic conditions and to verify the feasibility and validity of the model.

关 键 词：无人艇 任务决策 动态多实体贝叶斯网络 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]