基于LOF聚类的水下分布式协同定位技术  

Underwater distributed collaborative positioning technology based on LOF clustering

作　　者：胡鹏鹏 梁琴 赵显文 李捷[1] HU Pengpeng;LIANG Qin;ZHAO Xianwen;LI Jie(Sichuan Jiuzhou Electric Group Co.,Ltd.,Chengdu 610066,China)

机构地区：[1]四川九洲电器集团有限责任公司,四川成都610066

出　　处：《舰船科学技术》2025年第8期126-132,共7页Ship Science and Technology

基　　金：国家重点研发计划资助项目(2022YFE0136800)。

摘　　要：针对多节点分布式协同定位过程中,由于声呐系统存在误差、外部环境噪声和人为干扰等因素的影响,导致产生错误或较大偏差的测量数据,即离异值的问题,提出基于局部离群因子算法(LOF)聚类的分布式协同定位技术。该方法在分析定位结果分布特征的基础上,利用LOF聚类技术确定定位结果中的离群定位点,进而可以得到离群定位点所对应的分布式节点,然后选择合适阈值筛选存在问题的节点并将其剔除,最后将剩余分布式节点测量结果利用加权融合算法进行处理,完成最终定位。仿真实验证明,该技术有效降低了测距不准确对目标定位的影响,定位误差均值由1.33降低至0.32,有效提高了系统定位精度。A distributed collaborative positioning technology based on LOF clustering is proposed to address the issue of measurement data with errors or significant deviations,i.e.outliers,caused by inaccurate sonar systems,external environmental noise,and human interference in the process of multi node distributed collaborative positioning.On the basis of analyzing the distribution characteristics of localization results,this method uses Local Outlier Factor Algorithm(LOF) clustering technology to determine the outlier localization points in the localization results,and then obtains the distributed nodes corresponding to the outlier localization points.After selecting an appropriate threshold,the problematic nodes are screened and removed.Finally,the remaining distributed node measurement results are processed using a weighted fusion algorithm to complete the final localization.Through simulation experiments,it has been proven that this technology effectively reduces the impact of inaccurate ranging on target positioning error reduced from 1.33 to 0.32,effectively improving the systems positioning accuracy.

关 键 词：水下分布式定位 三边定位技术 LOF离群点检测 加权融合定位 

分 类 号：U675.79[交通运输工程—船舶及航道工程]