An improved multihop-based localization algorithm for wireless sensor network using learning approach

详细信息    查看全文

• 作者：Xiaoyong Yan^a ; ^b ; ^c ; ^{xiaoyong_yan@126.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Aiguo Song^a ; ^bAuthor Vitae ; Zhong Yang^cAuthor Vitae ; Wankou Yang^dAuthor Vitae
• 关键词：Wireless sensor network ; Multihop localization ; Kernel ridge regression
• 刊名：Computers & Electrical Engineering
• 出版年：2015
• 出版时间：November 2015
• 年：2015
• 卷：48
• 期：Complete
• 页码：247-257
• 全文大小：1382 K

文摘

Multihop range-free localization methods could obtain relatively reasonable location estimation in the isotropic network; however, during the practical application, it is often affected by various anisotropic factors such as the radio irregularity and barriers, which can significantly reduce its performance. In this paper, we propose a new approach for multihop localization in wireless sensor network based on nonlinear mapping and learning algorithm. The proposed method is simple, efficient, higher accuracy and no need to set complex parameter in that only hop-counts information and position information of the beacons are used for the localization. The proposed approach is composed of two steps: firstly, this algorithm uses kernel function to define the connectivity information (hop-counts) between nodes, then, learning method is used to guide and build the inter-node localization model; secondly, the hop-counts between the unknown nodes and beacons are used to estimate the coordinate of unknown nodes. We evaluate our algorithm under various anisotropic network and real environment, and analyze its performance. We also compare our approach with several available advanced approaches, and demonstrate the superior performance of our proposed algorithm in terms of location estimation adaptability.