الفهرس 
SummaryOnly 14 pages are availabe for public view
 |  | from | 175 |  |  |
| | | from | 175 | | |
Abstract
The use of Reconfigurable Receivers in cellular networks is a new architecture in wireless communications. With the introduction of Software Defined Radio (SDR) architecture, in which the basic radio functions are implemented in software rather than dedicated hardware, the cellular transceivers could be more flexible, more reconfigurable and intelligent. The explosion in the number of wireless standards in the last decade implies an increasing number of user terminals, offering a wide variety of user equipments for each wireless standard. Reconfigurable SDR-based receivers have presented intelligent multi-band multi-mode transceivers which can adopt to channel characteristics and user needs. Automatic Modulation Classification (AMC) is an essential part in any reconfigurable multi-band multi-mode SDR receiver. The automatic recognition of the modulation format of a sensed signal is considered as an intermediate step between signal detection and demodulation, and has an utmost importance in several commercial and military applications. AMC receivers can be used as an air interface identifier, it could be used in surveillance and spectrum management in Cognitive radios (CR), it enables a more efficient application of electronic counter measures such as jamming in warfare applications, and it’s also highly needed in networks deploying Adaptive Modulation Technique. The objective of this work is to study different reconfigurable receivers’ architectures that implement AMC. The design of AMC receiver architecture consists of two main stages: a pre-processing stage and a classifier stage. The pre-processing stage is distinct for every receiver as it depends entirely on the classifier used. Two different receiver architectures are presented in this work, based on the features required for classification. In the first part of the present study, reconfigurable receiver architecture for AMC based on the cyclostationarity characteristics of the modulated signals is presented. Stationary signals that undergo periodic transformations like modulation can be effectively modeled as cyclostationary signals, and has unique cyclic spectrums that are used for classification. This receiver applies a Neural Network to classify four different types of Digital modulation schemes: BPSK, QPSK, FSK and MSK. For higher order digital modulation format, such as: M-PSK and M-QAM, AMC receiver using clustering algorithms is presented in the second part of the thesis. A set of pre-processing stages is introduced, and three different clustering algorithms are compared along with the corresponding clustering validity measures.