EFFICIENT LOCALIZATION ALGORITHMS FOR WIRELESS SENSOR NETWORKS

Abstract

In this dissertation, efficient localization algorithms for wireless sensor networks are represented. Localization algorithms are widely used in commercial systems and application. The localization techniques are anticipated to be developed for various environments and reduce the localization error for accurate location information because the user demands for more accurate positioning systems for medical care, home networks, and monitoring applications in personal range environments. A well-known localization system is GPS, with applications such as mobile navigation. The GPS shows good performance on road or roughly finding location system in outdoor environments but limited in indoor environments. Due to the development of handsets like smart phone, the users can easily receive the GPS signals and other RF signals including 3G/4G/5G signals, WLAN (Wireless Local Area Networks) signals, and the signals from other sensors. Thus, the various systems using localization schemes are developed, especially, the WSNs (Wireless Sensor Networks) localization system is actively studied in indoor environment without GPS. In this dissertation, the range-free localization algorithm and the range-based localization algorithm are reported for WSNs localization system. The range-free localization algorithms are proposed before to estimate location using signal database, called signal map, or the anchor nodes of antenna patterns, or ID configuration of the linked anchor nodes, etc. These algorithms generally need to additional hardware or have low accuracy due to low information for location estimation. The range-based algorithms, equal to distance-based algorithms, are based on received signal strength, RSSI, or time delay, TOA and TDOA, between the anchor nodes and a target node. Although the TOA and TDOA are very accurate distance estimation schemes, these scheme have the critical problem, the time synchronization. Although RSSI is very simple to setup the localization system with tiny sensors, the signal variation causes severe distance estimation error. The angle estimation, AOA, provides additional information to estimation the location. However, AOA needs additional hardware, the antenna arrays, which is not suitable for tiny sensors. In this dissertation, range-free and range-based localization algorithms are analyzed and summarized for WSNs with tiny sensors. The WSNs localization systems are generally used range-based algorithm. The range-based algorithms have major source of distance estimation error, and the distance estimation error causes severe localization error. In this dissertation, the localization error mitigation algorithms are proposed in two dimensional environments and three dimensional environments for WSNs. The mitigation algorithms in two dimensional environments consist of several steps, which are distance error mitigation algorithm, location error mitigation algorithm, and bad condition detection algorithm. The each algorithm is effective to reduce the localization error, but the accuracy of location estimation is the best when they are combined. The performance of proposed algorithms is examined with variation of received signal strength and it is confirmed that the combined proposed algorithm has the best performance rather than that of conventional scheme and each proposed algorithms. The three dimensional localization uses Herons formula of tetrahedron to calculate the target height, then transforms a two dimensional location computed by LLSE into a three dimensional estimated location. Simulation results validate the accuracy of the proposed scheme.