基于LMS-KF算法的室内移动目标定位  

Indoor positioning of moving target based on LMS-KF algorithm

作　　者：程谟方 董元方[2] 姜淑华[1] 

机构地区：[1]长春理工大学电子信息工程学院,吉林长春130022 [2]长春理工大学经济管理学院,吉林长春130022

出　　处：《计算机工程与设计》2016年第7期1927-1931,共5页Computer Engineering and Design

基　　金：国家级大学生创新创业基金项目(2014S043);吉林省科技发展计划基金项目(20140101199JC)

摘　　要：针对时变环境下室内定位精度较差的问题,提出一种最小均方误差-卡尔曼滤波(LMS-KF)定位算法。应用最小二乘法拟合离线状态下的室内信号传播模型;在在线阶段,针对不断变化的室内环境,采用最小均方误差自适应滤波实时调节传播模型,应用三边定位法进行定位;将定位结果作为观测值,应用卡尔曼滤波对定位结果进行优化。仿真结果表明,在时变的室内环境下,LMS-KF算法能有效提高室内定位精度,使定位误差在1.5m以上的概率为17.5%,比未校正时减小了28.5%,比应用传统KF算法减小了12.5%。To overcome the problem of poor indoor positioning accuracy in time-varying environment,the least mean square-Kal-man filter algorithm was proposed.The curve of the indoor signal propagation model was fitted using the least square method in the offline state.In view of changing indoor environment,the propagation model was adj usted in real time using the adaptive fil-ter method of least mean square,and then the trilateral positioning method was applied for locating.The positioning results were taken as observations and it was optimized using the method of Kalman filter.The simulation results show that the LMS-KF （least mean square-Kalman filter）algorithm can effectively improve the indoor positioning accuracy in time-varying environment. The positioning error probability in 1.5 m or more is 17.5% which is decreased by 28.5% and 12.5% respectively,compared with the locating algorithms without calibration and KF （Kalman filter）algorithm.

关 键 词：时变环境 信号强度 最小二乘法 最小均方误差滤波 卡尔曼滤波 

分 类 号：TP272[自动化与计算机技术—检测技术与自动化装置]