基于凸优化方法的室内NLOS误差抑制算法  

Indoor NLOS Error Mitigation Algorithm Based on Convex Optimization Method

作　　者：张龙[1] 任修坤[1] 王盛[1] 张伟[1] ZHANG Long;REN Xiukun;WANG Sheng;ZHANG Wei(Information Engineering University, Zhengzhou 450001, China)

机构地区：[1]信息工程大学,河南郑州450001

出　　处：《信息工程大学学报》2020年第3期279-284,共6页Journal of Information Engineering University

基　　金：国家自然科学基金资助项目(61272041)。

摘　　要：非视距(non-line-of-sight,NLOS)误差是导致室内定位精度低、稳定性差的一个重要原因,现有NLOS误差抑制算法存在复杂度较高、鲁棒性较差等问题。提出一种基于凸优化方法的室内NLOS误差抑制算法,为保证定位鲁棒性,该算法先给出鲁棒最小二乘(robust least squares,RLS)形式的位置估计问题,再依据遮挡情况不同,将定位环境分为轻微遮挡环境和严重遮挡环境,并根据两种环境NLOS误差特性,引入新的松弛条件,将上述位置估计问题分别转化为二次约束二次规划问题和二阶锥规划问题并求解。仿真实验表明,相比已有算法,在不同应用场景下,所提算法提高了定位精度,并且有效降低了无解个数,增强了鲁棒性。The non-line-of-sight(NLOS)error is an important cause of low indoor localization accuracy and poor stability.The existing NLOS error mitigation algorithms have problems of high complexity or poor robustness.To this end,this paper proposes an NLOS error suppression algorithm based on convex optimization method.The algorithm firstly gives the position estimation problem in the form of robust least squares(RLS).Secondly,according to the seriousness of the occlusion of the positioning scene,the scene is divided into two categories.Combining the NLOS error characteristics of the two scenarios,a new relaxation condition is introduced.The above localization estimation problem is transformed into a quadratic constrained quadratic programming problem(QCQP)and a second-order cone programming problem(SOCP)and solved.Simulation experiments show that compared with the error suppression contrast algorithm,the algorithm improves the localization accuracy,and effectively reduces the number of unsolved,and enhances the robustness of the algorithm.

关 键 词：非视距 鲁棒最小二乘 凸优化 二次约束二次规划 二阶锥规划 

分 类 号：TN97[电子电信—信号与信息处理]