Ph.D. Tezi Görüntüleme | |||||||||||||||||||||
|
|
||||||||||||||||||||
| Summary: Two of the most important physical layer problems of digital communication systems, the convergence speed and performance of adaptive training algorithms are subjected in this thesis. Although digital communication systems have been advanced in all senses, the speed of channel parameter estimation algorithms and its accuracy are found the main limitations forming the physical layer of the communication standards. Therefore, in an attempt to improve the performance and convergence speed of adaptive training algorithms, this thesis is dedicated to implement a fuzzy outer loop controller in order to control the error trajectories of adaptive training algorithms. Two important algorithms are chosen for this reason. These are the conventional LMS and the well known blind technique, the CMA, algorithms, since both algorithms suffer from slow convergence. The essential aim of the thesis has been achieved when the performance of the LMS is increased becoming comparable to whose using the RLS algorithm, and also the convergence features of the blind CMA algorithm has also been significantly improved with the proposed fuzzy outer loop controller. Thus, the thesis proposes an experience based fuzzy outer loop controller for the stochastic signal processing tools of communications, in order to increase the convergence speed, control the error function and decrease the error floor. Therefore, the major conventional channel and equalizer estimation methods are studied and compared with the LMS when its step size parameter is controlled by the proposed fuzzy outer loop controller. For performance comparisons, the HIPERLAN/1 (QPSK) and HIPERLAN/2 (OFDM) standards are used as the working platforms during simulations. Secondly the thesis concentrates on the major blind training techniques and their performances are compared with the proposed Fuzzy-CMA algorithm. The required computational complexity of the outer loop controller is not higher than the simple LMS algorithm. The thesis also includes the subjects’ adaptive training while carrier frequency offset (CFO) and Doppler presents in the channel, which are well known features of wireless communications.Keywords: Channel Estimation, Channel Equalization, Blind Channel Equalization, CMF-DFE, MMSE-DFE, Fuzzy Outer Loop Controller, CFO, Doppler, Computational Complexity | |||||||||||||||||||||