Heterogeneous Wireless Systems Based Seamless Positioning Technologies

Navigation and positioning are emerging growth aspects of wireless communication service, and seamless localization is the development trend of positioning technology. To meet the increasing public requirements Location Based Services(LBS), multiple existing radio localization technologies should be combined for the purpose of better performance.Nowadays,outdoor environment positioning technology is well-developed. Global Navigation Satellite System(GNSS) based LBS could meet the demand for outdoor navigation of ordinary people. But in complex environments, such as indoor and urban canyon scenarios, GNSS is invalid. Indoor scenario was once supposed to be the blind area of positioning service. Thus, the realization of indoor location was once regarded as the bottleneck to realize seamless positioning. However, recently indoor localization has become a research hot spot. A variety of indoor positioning techniques have been proposed, which provide plenty of technology choices for indoor LBS, and some of them are already at commercial attempt. But indoor localization and outdoor localizat ion are developed independently, and the gap between them still exists, which blocks the implementation of seamless LBS. Therefore, the elimination of this gap has become a new hot research item of seamless positioning technology, and it is now the last bottleneck of seamless LBS to design a positioning system, which works continuously and smoothly between indoor and outdoor scenarios.This thesis analyses the current research works of indoor and outdoor localization techniques, particularly the developing process of seamless positioning. Based on the development statuses of existing positioning techniques, the development directions of seamless positioning technology are summarized. Among them, fusing heterogeneous localization systems is selected as the research direction of this paper, in order to maximize the utilization of existing positioning resources. The main topics of this research are as follows, i.e., solve the problem of radio map establishing in fingerprint positioning, design handover algorithms for heterogeneous seamless localization systems, and enlarge the application scope of fingerprint positioningto outdoor area to provides seamless localization service in city canyon scenarios.Then, related theories are also presented. The operation mechanism of fingerprint positioning is introduced, together with the basic principles of radio map fast establish technology. The common means of handover algorithms are also summarized as well.As indoor localization has not been completely researched yet, practical indoor positioning system is sill not available. This thesis firstly focuses on blind area in traditional navigation methods. Fingerprint positioning technique is regard as a promising choice for indoor localization, because it utilizes the complexity of indoor environment. This thesis proposes a self-growing approach on radio map construction. Data from volunteers are utilized to enlarge the coverage of the radio map. The proposal aims to solve the problem of radio map establishing cost, which is the most dominant restriction of fingerprint positioning technology. For the first time, KNN algorithm is used as the radio map recover method, which is the core technology of participatory perception based radio map establish technology. Compared with different recover algorithms, KNN algorithm has obvious advantages.Further, this thesis analyses the handover problem in positioning systems, which is important to bridging different systems. Based on the feature of positioning systems, handover issue is divided into handover multiple items, such as designing handover zone, handover trigger and judgement etc. An accuracy estimation algorithm and an Euclidean distance based handover algorithm are presented for fingerprint positioning systems, which are used as the inter-system handover and system edge handover method respectively.Meanwhile, to enhance the robustness and extend the application scenarios of fingerprint positioning system, this thesis analyses the challenges of fingerprint positioning systems in an open scenario. It is found that sporadic large scale positioning error caused by environment changing is the main challenge of this issue. An Adaptive Feature Space(AFS) algorithm and a movement feature constraint tracking algorithm are proposed to avoid this problem. Both of the aforementioned algorithms are verified in practical systems.