《计算机应用研究》|Application Research of Computers

基于CMAES-SVR的WLAN室内定位算法研究

Research on WLAN indoor positioning algorithm based on CMAES-SVR

免费全文下载 (已被下载 次)  
获取PDF全文
作者 饶华,王忠,李欣
机构 四川大学 电气工程学院,成都 610065
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2019)08-058-2514-04
DOI 10.19734/j.issn.1001-3695.2018.07.0408
摘要 针对传统WLAN指纹定位算法中存在的定位精度低、稳定性差、实时性不高等问题,提出一种基于CMAES-SVR的WLAN室内定位算法。该算法首先对接入点(AP)的接收信号强度(RSS)进行统计分析,采用高斯滤波对信号进行预处理,然后利用K-means聚类算法将原始指纹数据库中的定位区域进行聚类分块;其次采用协方差矩阵自适应进化策略(CMAES)优化支持向量回归机(SVR)参数,从而建立CMAES-SVR室内定位学习模型,通过该模型分别构建各定位子区域中RSS信号与物理位置非线性映射关系;最后判断测试点所属类簇,根据该类簇中训练好的CMAES-SVR模型进行回归预测。实验结果表明,与WKNN、传统SVR以及PSO-SVR算法相比,该算法在定位精度、稳定性以及实时性方面均有所提高。
关键词 室内定位; 位置指纹; 聚类分析; 协方差矩阵自适应进化策略; 支持向量回归
基金项目 四川省科技厅科技支撑计划项目(2015FZ061)
本文URL http://www.arocmag.com/article/01-2019-08-058.html
英文标题 Research on WLAN indoor positioning algorithm based on CMAES-SVR
作者英文名 Rao Hua, Wang Zhong, Li Xin
机构英文名 College of Electrical Engineering,Sichuan University,Chengdu 610065,China
英文摘要 Aiming at the problems of low location accuracy, poor stability and low real-time performance in traditional WLAN fingerprint positioning algorithm, this paper proposed a WLAN indoor positioning algorithm based on CMAES-SVR. Firstly, the algorithm statistically analyzed the received signal strength(RSS) of access point(AP) and used the Gaussian filter to pretreat the RSS signal. It adopted the K-means clustering algorithm to partition the location area of original fingerprint database. Secondly, the covariance matrix adaptive evolutionary strategy(CMAES) optimized the support vector regression(SVR) parameters so as to establish the CMAES-SVR indoor positioning learning model. This model constructed the nonlinear mapping relationship between RSS signal and physical location in each sub location. Finally, the algorithm determined the cluster of the test points and carried out the regression prediction according to the trained CMAES-SVR model. The experimental results show that compared with WKNN, traditional SVR and PSO-SVR algorithm, the algorithm improves location accuracy, stability and real-time performance.
英文关键词 indoor position; location fingerprint; cluster analysis; CMAES; SVR
参考文献 查看稿件参考文献
 
收稿日期 2018/7/3
修回日期 2018/8/17
页码 2514-2517,2521
中图分类号 TP393
文献标志码 A