摘要
为解决最小二乘支持向量回归(least-square support vector regression,LS-SVR)定位精度不高的问题,提出基于LS-SVR的混合定位算法,充分考虑未知节点之间的距离信息在定位过程中的有效修正作用。通过LS-SVR算法提供初始值,提高多元Taylor级数展开法的收敛速度;通过多元Taylor级数展开法,充分利用未知节点之间的距离信息,减小测距误差造成的定位误差。仿真结果表明,与传统LS-SVR定位算法相比,混合定位算法的精度更高,减少了正则化参数和核参数的选取对定位精度的影响。
To solve the problem of low positioning accuracy of least-square support vector regression(LS-SVR),a hybrid positioning algorithm based on LS-SVR was proposed,which considered effective correction of the distance information between unknown nodes in the positioning process.The initial values obtained by LS-SVR algorithm can improve the convergence speed of multivariable Taylor series expansion method,and the multivariate Taylor series expansion method can reduce the position error caused by ranging error through making full use of the distance between the unknown node information.Experimental results indicate that the proposed algorithm can improve positioning accuracy and reduce the sel
作者
夏斌
梁春燕
袁文浩
谢楠
XIA Bin;LIANG Chun-yan;YUAN Wen-hao;XIE Nan(College of Computer Science and Technology,Shandong University of Technology,Zibo 255000,China)
出处
《计算机工程与设计》
北大核心
2018年第11期3318-3321,3339,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61701286
11704229)
山东省自然科学基金项目(ZR2017MF047
ZR2015FL003)
关键词
多元泰勒级数展开
定位模型
最小二乘支持向量回归
定位精度
混合算法
multivariable Taylor series expansion
positioning model
least-square support vector regression
positioning accuracy
hybrid algorithm
