Abstract
In wireless networks, the location of a mobile user (MU) is a vital piece of information for their communication and other services. The impressive increase in the use of mobile handsets and location based services (LBSs) in recent years has shown the increasing demand of localizing MUs continuously to have ubiquitous and pervasive computing. Global positioning system (GPS) has been accepted as the localization technique for outdoor environments. However, GPS cannot be used for indoor localization due to signal blocking in most of indoor environments. This limitation has shown the need for indoor localization system. Many techniques have been developed to provide location information in indoor environments. One of the available technologies for indoor localization is the use of radio frequency (RF) signal. Among this technology WLAN-based indoor localization have been attracting more applications. In these contents, received signal strength (RSS)-based fingerprinting technique is the preferable technique for indoor positioning. However, the key implementation challenge of this technique is the need of site survey for the construction of radio map, which is time consuming, labour intensive and not exible. Motivated by these, the major contributions of this thesis are in threefold: 1. A novel approach for reducing the calibration effort of the radio map construction for any indoor environment is proposed. The basic idea of the proposed approach is to triangulate the surveyed calibration points according to the principle of Delaunay triangulation. Then, the complete radio map is formed by adopting cubic interpolation inside each triangle. The experimental results show that the proposed approach can reduce the effort of calibration while providing comparable localization accuracy with the conventional fingerprinting technique. Additionally, it outperforms the existing RSS-interpolation approaches in terms of localization accuracy and system precision. 2. A novel non-sensor-aided crowdsourcing approach to reduce the calibration efforts of the construction of corridors radio map is proposed. This approach is based on surveying some calibration points and benefit from MUs' movement inside the building to construct the complete radio map. The system administrator surveys both ends of each corridor in the building. Then, by using the collected data from the MUs during their movement, the RSS at different points along the corridors can be estimated, i.e., the corridors radio map is built. Recently collected RSS data are frequently used to adapt the radio map to the environmental changes. The experimental results show that the proposed approach can be used to build corridors radio map with high accuracy and achieve localization accuracy comparable with conventional fingerprint. In addition, the proposed approach provides comparable results with current approaches for reducing the calibration effort and outperforms them in some aspects. 3. The effect of using crowdsourcing approach for fingerprint-based indoor localization systems on localization accuracy is studied by deriving the Cramer-Rao lower bound (CRLB) for this case. The analytical results indicate that using crowdsourcing approach reduces the localization accuracy. However, the results also suggest that increasing the number of access points (APs) will minimize this effect.