For the past few years, there has been a high interest in mobile terminal (MT) positioning. The primary motivation for the development of mobile positioning systems was due to the mandatory requirement of E-911 service by the U.S. FCC. Although the starting was because of security-emergency need, later it has found various applications in many fields, e.g., to increase data throughput in cellular systems.
There exist many algorithms developed for the MT localization problem. The traditional geometrical localization methods are designed to work under line-of-sight (LoS) conditions. However, LoS conditions might not always be present between the base station (BS) and the MT. Therefore, fingerprinting-based localization techniques which are also the subject of this thesis attract attention because of their ability to work also in multipath and non-line-of-sight (NLoS) environments.
In this thesis, we introduce new fingerprinting algorithms, namely power delay Doppler-profile fingerprinting (PDDP-F) algorithms exploiting the mobility of the MT. The purpose is to increase the localization accuracy by utilizing the Doppler dimension. We also investigate the localization performances of power delay-profile fingerprinting (PDP-F) and PDDP-F algorithms via the derivation of Cramer-Rao bounds (CRBs). The impact of the network geometry is also studied.
Another subject we deal with is the pairwise error probability (PEP) analysis for PDP-F methods. PEP is a well-known notion in digital communications, and we import it into the field of localization to derive the probability of making a decision in favor of a wrong position.
The last topic we work on is about MT tracking based on adaptive Kalman filtering. Various mobility models are compared in terms of their position prediction errors.
Systèmes de Communication
© EURECOM. Personal use of this material is permitted. The definitive version of this paper was published in Thesis and is available at :