Performance evaluation of Location fingerprinting techniques in WLAN systems

Abstract

The research focuses on location fingerprinting as one practical and popular localization technique due to its use of the available wireless infrastructure (WLAN). The technique is composed of two stages namely offline and online. In the offline stage, fingerprints are constructed from the measured RF signals across a grid, in a designated indoor (floor/office) area, and stored into a database. In the online stage, a mobile terminal can be located (position estimation) by constructing an online fingerprint and comparing it with the offline database of fingerprints through well-known pattern recognition techniques. Received Signal Strength (RSS) is one of the earlier and simplest fingerprinting techniques used and readily available in IEEE 802.11 standards implementation. However, because the RSS fingerprint provides a weak fingerprint structure, other RF parameters such as Channel Impulse Response (CIR), frequencydomain Channel Transfer Function (CTF) and autocorrelation of the CTF or Frequency Coherence Function (FCF), are more complex but unique, have been proposed as alternative fingerprints. The research focuses on developing an empirical framework for comparing the performance of different fingerprint structures such as RSS, CIR, CTF and FCF which is essential to assess the strengths and weaknesses of those techniques. The work is divided into two parts; in the first part a simulation framework has been developed using the Ray-Tracing software in order to emulate the indoor wireless channel. The CIR-based simulation results showed noticeable improvements in the location estimation compared to the RSS-based approach. The simulation framework also highlights new findings for some parameters that affect the performance of the CIR-based fingerprinting with respect to the system bandwidth and training point spacing. In the second part a frequency domain channel measurements in the IEEE 802.11 band are used to evaluate the accuracy of the fingerprints and their robustness to human induced motion perturbations of the channel. Localization performance is analysed and explained using the spatial and temporal radio propagation characteristics. Specifically the concept of coherence region is introduced to explain the spatial properties of the fingerprints while the temporal properties are explained by studying the impact of the Doppler spread in time-varying channels on the time coherence of the RF fingerprint structures. FCF fingerprint structure showed performance in the spatial and temporal domains.

Date of Award	2013
Original language	American English
Supervisor	Saleh Al-Araji (Supervisor)