Indoor localization techniques can make many indoor applications more intelligent and bring convenience to our work and life. Wireless sensor network provides a new solution for indoor localization and has the advantages of small sensor nodes, low power consumption and easy deployment. This paper focuses on the research of wireless sensor network based indoor localization techniques with support from the National High Technology Research and Development Program of China and the National Natural Science Foundation of China. The content involves the implementation of wireless sensor based indoor localization system, analysis of indoor signal characteristics, localization of static nodes, localization of reference nodes and human tracking. The work of this paper is summarized as follows. 1. A wireless sensor network based indoor localization system is designed. It consists of a wireless sensor network and a monitoring network, which form a hierarchical structure. Sensor nodes are compatible with IEEE 802.15.4 standard and are connected together with Collection protocol. The monitoring network, which is a LAN (Local Area Network), is used to manage data from several wireless sensor networks. The localization system can be easily expanded and applied in many different indoor environments. 2. Empirical signal models in environments with and without human are established and effects of many factors on radio signal strength are analyzed. In indoor environment, the propagation of wireless signal is affected by many factors, which makes it hard to derive precise theoretical signal models. To solve this problem, empirical signal models are built. Experiments are also conducted to analyze the effects of sensor nodes’ height, signal frequency, sending power, antenna orientation and distance from wall etc. on radio signal strength. RSS based ranging method is also discussed. 3. Two radio fingerprint based localization algorithms, improved radio fingerprint matching algorithm and radio fingerprint based weighted centroid localization algorithm, are proposed. Analysis on traditional radio fingerprint matching algorithm suggests that it can get the maximum likelihood estimation of unknown nodes’ positions. But traditional radio fingerprint matching algorithm requires large computation resource which is unavailable in wireless sensor network. To reduce the complexity of localization, improved radio fingerprint matching algorithm and radio fingerprint based weighted ce...
修改评论