An Efficient Nonparametric Belief Propagation-Based Cooperative Localization Scheme for Mobile Ad Hoc Networks
In mobile ad hoc networks, nonparametric belief propagation (NBP) algorithm is a promising cooperative localization scheme because of high accuracy, applicability to non-Gaussian uncertainty. However, the high computational cost limits the application of
- PDF / 337,791 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 4 Downloads / 179 Views
stract. In mobile ad hoc networks, nonparametric belief propagation (NBP) algorithm is a promising cooperative localization scheme because of high accuracy, applicability to non-Gaussian uncertainty. However, the high computational cost limits the application of NBP. To solve the problem, an efficient and practical NBP-based cooperative localization scheme is proposed. In the scheme, the issues of anchor node selection, node mobility and non-Gaussian uncertainty are considered. Firstly, anchor nodes are selected based on a distributively clustered network. Then the cooperative localization process is performed, in which a practical ranging error model is employed. Moreover, to mitigate the influence of node mobility, the re-selection process of anchor nodes is conducted when necessary. The simulation results demonstrate the efficiency of the proposed scheme in improving the positioning accuracy and reducing the computational cost compared with the conventional NBP method. Keywords: Mobile ad hoc networks Ranging error model
1
·
NBP
·
Anchor node selection
·
Introduction
In mobile ad hoc networks (MANETs), accurate positioning information is crucial since it enables a wide variety of applications, such as emergency services, first responders operations and factory automation [1,2]. In typical localization schemes, nodes in a MANET can be divided into anchor nodes, which have known positions and account for a small proportion in the nodes, and agent nodes that need to be located by utilizing the information from anchor nodes. Generally, the existing range-based localization schemes can be classified into non-cooperative schemes and cooperative schemes. In the non-cooperative schemes [1], an agent node is located only depending on the measured distances with neighboring anchor nodes. For the cooperative schemes, by contrast, agent nodes estimate their positions through ranging and exchanging information with neighboring nodes, including anchor nodes and other agent nodes. Cooperation c ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2018 Q. Chen et al. (Eds.): ChinaCom 2016, Part II, LNICST 210, pp. 519–528, 2018. DOI: 10.1007/978-3-319-66628-0 49
520
C. Xu et al.
among the agent nodes is highly beneficial for improving the performance of localization processes on accuracy and coverage [1]. Lately, extensive works have been focused on cooperative localization [3–7], and most of these schemes are based on belief propagation (BP) algorithm and its extension algorithms [3,6] for high accuracy and distributed implementation. BP is an efficient messagepassing method of estimating the a posterior marginal probability density function (PDF) for the positions of the agent nodes in the network, but the inability in resolving non-Gaussian uncertainty, which is a common occurrence in practical localization scenarios, limits the application of BP. As an extension of BP, nonparametric belief propagation (NBP) algorithm [3] is sample-based and can be applied in the positioning systems w
Data Loading...