Hybrid and intelligent positioning techniques in heterogeneous and cognitive networks -

Abstract

Nowadays, the availability of the location information becomes a key factor in today’s communications systems for allowing new location based services. In outdoor scenarios, the Mobile Terminal (MT) position is obtained with high accuracy thanks to the Global Positioning System (GPS) or to the standalone cellular systems. However, the main problem of GPS or cellular systems resides in the indoor environment and in scenarios with deep shadowing effect where the satellite or cellular signals are broken. This thesis is divided into two main parts. In the first part, we present a potentially good candidate for critical positioning scenarios with the lack of hearability between the Unlocated Mobile Terminal (UMT) and the Anchor Nodes (AN). Indeed, in many cases, only one or two ANs are communicating with the UMT. The proposed positioning algorithm is based on hybrid data fusion and its extension to the tracking step by adopting the Minimum Entropy Criterion to diminish the shortcoming of the Unscented Kalman Filter and Particle Filter, usually used in this field. The proposed solution is divided into two phases: the learning phase and the processing phase. Using Radial Basis Functions, the learning phase allows an accurate model of the probability density function of the positioning error while the processing phase aims at reducing the estimation error. We show that the proposed algorithm reaches an accuracy of 1m squared in terms of Mean Square Error (MSE). Not only localizing MTs, but also localizing multiple transmitters in a region is also another objective tackled in this thesis. Even though this problem is applicable in different applications, the most prominent one is the cognitive radio context. We are interested in the uncoordinated system where the cognitive node operates in an opportunistic manner. In order to avoid interference, the cognitive system is responsible to recognize the area where there are active primary users. Assuming the location of primary users and their activity are not known, we