一种基于核岭回归与卡尔曼滤波的定位算法  

作　　者：康颜爽 苏胜君[1] 施伟斌[1] 乐燕芬[1] 韩承毅 KANG Yan-shuang;SU Sheng-jun;SHI Wei-bing;LE Yan-fen;HAN Cheng-yi(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《软件导刊》2020年第9期74-77,共4页Software Guide

基　　金：国家自然科学基金青年项目(51705324)。

摘　　要：传感器定位技术作为无线传感器网络的重要课题之一,为实现目标实时定位,提出一种基于核岭回归与卡尔曼滤波的定位算法。该算法在离线阶段使用核岭回归算法(KRR)对无线位置指纹数据库进行训练,从而得到一个可反映信号强度指标(RSSI)与位置坐标之间映射关系的函数;在线阶段先利用离线阶段得到的函数对目标进行粗定位,再结合卡尔曼滤波(KF)方法对目标进行精确定位。实验结果表明,在真实室内办公环境下,相比KNN算法与核函数(Kernel)算法,该算法能实现更好的定位精度,平均定位误差为1.898 3m。Sensor location is one of the important topic in wireless sensor networks. In order to achieve real-time location of the target,a localization algorithm based on Kernel Ridge Regression(KRR)and Kalman Filter(KF)is proposed in this paper. In the off-line phase,KRR algorithm is used to train the wireless location fingerprint database so as to obtain a function that could reflect the relationship between received signal strength indicator(RSSI)and position coordinates. In the online tracking stage,the function is obtained in the off-line phase to roughly estimate the position,then the KF is used to accurately locate the target. The experimental results show that in the real indoor office environment,compared with KNN and Kernel algorithm,the algorithm proposed in this paper can achieve better positioning accuracy,and the algorithm average error is 1.898 3 m.

关 键 词：目标定位 RSSI 核岭回归 无线位置指纹数据库 卡尔曼滤波 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]