基于EBM的UUV对目标威胁评估方法  

Target Threat Assessment Method for UUVs Based on EBM

作　　者：刘书伟 程健庆[1] 刘凯[1] LIU Shuwei;CHENG Jianqing;LIU Kai(Jiangsu Automation Research Institute,Lianyungang 222006,China)

机构地区：[1]江苏自动化研究所,江苏连云港222006

出　　处：《水下无人系统学报》2025年第1期164-172,共9页Journal of Unmanned Undersea Systems

摘　　要：针对传统的目标威胁评估方法处理复杂战场态势数据时,缺乏数据挖掘能力和神经网络算法解释性不足等问题,提出基于可解释增强机(EBM)的无人水下航行器(UUV)对目标威胁评估模型。EBM作为一种先进的机器学习技术,巧妙融合了梯度提升与广义加性模型,实现了线性模型的高可解释性与梯度提升算法准确性的完美结合。文中对EBM模型的性能进行了全面评估,并与分类提升、自适应提升以及深度学习等几种主流机器学习方法进行了比较。通过仿真实验发现, EBM模型在保持高可解释性的同时,对威胁等级识别的准确度也高达98.10%。这一结果不仅验证了EBM模型在复杂战场态势分析中的有效性,也为UUV的自主决策提供了坚实的理论基础和技术支持。In order to solve the problems of lack of data mining ability and insufficient explanatory nature of neural networkalgorithms when traditional target threat assessment methods process complex battlefield situation data,this paper proposed athreat assessment model for unmanned undersea vehicles(UUVs)based on explainable boosting machine(EBM).As anadvanced machine learning technology,EBM cleverly integrates gradient boosting and generalized additive model to achievethe perfect combination of high interpretability of linear model and accuracy of gradient boosting algorithm.In this paper,theperformance of the EBM model was comprehensively evaluated and compared with several other mainstream machinelearning methods,including classification enhancement,adaptive enhancement,and deep learning.The simulation experimentsfind that the EBM model not only maintains high interpretability but also has an accuracy of 98.10%in the identification ofthreat levels.This result verifies the effectiveness of the EBM model in complex battlefield situation analysis and provides asolid theoretical foundation and technical support for UUV’s autonomous decision-making.

关 键 词：威胁评估 无人水下航行器 可解释增强机 梯度提升 

分 类 号：TJ630[兵器科学与技术—武器系统与运用工程] U663[交通运输工程—船舶及航道工程]