摘要
We propose a method to improve positioning accuracy while reducing energy consumption in an indoor Wireless Local Area Network(WLAN) environment.First,we intelligently and jointly select the subset of Access Points(APs) used in positioning via Maximum Mutual Information(MMI) criterion.Second,we propose Orthogonal Locality Preserving Projection(OLPP) to reduce the redundancy among selected APs.OLPP effectively extracts the intrinsic location features in situations where previous linear signal projection techniques failed to do,while maintaining computational efficiency.Third,we show that the combination of AP selection and OLPP simultaneously exploits their complementary advantages while avoiding the drawbacks.Experimental results indicate that,compared with the widely used weighted K-nearest neighbor and maximum likelihood estimation method,the proposed method leads to 21.8%(0.49 m) positioning accuracy improvement,while decreasing the computation cost by 65.4%.
We propose a method to improve positioning accuracy while reducing energy consumption in an indoor Wireless Local Area Network （WLAN） environment. First, we intelligently and jointly select the subset of Access Points （APs） used in positioning via Maximum Mutual Information （MMI） criterion. Second, we propose Orthogonal Locality Preserving Projection （OLPP） to reduce the redundancy among selected APs. OLPP effectively extracts the intrinsic location features in situations where previous linear signal projection techniques failed to do, while maintaining computational efficiency. Third, we show that the combination of AP selection and OLPP simultaneously exploits their complementary advantages while avoiding the drawbacks. Experimental results indicate that, compared with the widely used weighted K-nearest neighbor and maximum likelihood estimation method, the pro- posed method leads to 21.8% （0.49 m） positioning accuracy improvement, while decreasing the computation cost by 65.4%.
基金
the High-Tech Research and Development Program of China，the National Seience Foundation for Young Scientists of China，the China Postdoctoral Science Foundation funded project
作者简介
Deng Zhian,Ph.D.Student at the School of Electronics and Infonnation Engineering,Harbin Institute of Technology,China.His current research interests include wireless positioning,machine learning and signal processing.Email:dengzhianan@163.com;Xu Yubin,professor and doctoral supervisor at the School of Electronic s and Information Engineering,Harbin Institute of Technology,China.Iris main research interests are navigation and positioning,communication networks and private mobile corn-munications.Email:ybxu@hit.edu.cn;Ma Lin,lecturer at the School of Electronics and Information Engineering,Harbin Institute of Technology,China.His main research interests are wireless positioning and wideband wireless communi-cations.Email:malin@hit.edu.cn.
