Research And Implementation Of TDOA-Based Wireless LAN Localization System

The dramatic advance in wireless communication technologies has triggered the wide spread of Location-based Services(LBS) in all aspects of people’s lives. Localization technology is the foundation of LBS. Since currently prevalent satellite aided localization systems cannot work well in built-up and indoor scenarios, performing localization with the ubiquitous Wireless Local Area Network(WLAN) is beneficial to expand the scope of LBS applications, and thus has important research and practical values. In WLAN localization, Time Difference of Arrival(TDOA) technique, which does not require the synchronization between transmitters and receivers, has attracted more and more attention.This thesis investigates the issue of TDOA based WLAN localization. In order to combat the effect of low timestamp resolution in conventional frame detection, it proposes the correlation-peak-amendment based fine timestamp estimation method as well as the multi-beacon-averaging based super-resolution timestamp estimation method, and effectively improves the localization accuracy. Moreover, in order to reduce the complexity of system implementation, it considers the open and reconfigurable characteristics of software radio, and designs a Universal Software Radio Peripheral(USRP) and GNURadio based system implementation scheme, which significantly cuts down the system development cost and period. Furthermore, according to the research above, this thesis completes a demo system for TDOA based WLAN localization. The demo system comprises a receiver and a system backend. The receiver, developed based on USRP, is used to capture beacon frame and record the beacon arrival time(Rx timestamp). The system backend is made up by localization client and localization server. The former, developed based on GNURadio, can extract Rx timestamp from the bit stream and forward it to the latter; the latter, built on a PC, is responsible for gathering and saving the Rx timestamps of multiple receivers, and calculating the target position according to the least mean square error algorithm.Finally, this thesis gives some field tests on timestamp accuracy and system localization performance. Timestamp accuracy test results verify the validity of improved timestamp estimation methods proposed in this thesis. Localization performance test results show that, the localization error of TDOA based WLAN localization system designed and completed by this thesis is below 1m as expected.