Improved synchronization in the mobile communications environment
Abstract
Some new carrier synchronization architectures for use in the time-varying, frequency nonselective Rayleigh fading channel are introduced. The objective of the algorithms developed is to provide improved performing and more spectrally efficient carrier synchronization compared to the standard transmitted reference techniques in the mobile communications environment. The architectures are based on approximations to the optimal detection scheme. The performance of the algorithms presented are characterized by Monte Carlo simulation and compared to ideal coherent detection and standard pilot symbol assisted modulation (PSAM) detection. Some diversity combining algorithms are also developed which provide significantly better bit error probability (BEP) performance in fading. Some new theoretical BEP curves for coherent detection of 16-QAM and 64-QAM modulations in fading for various demodulation schemes are also presented in this work. A crude approximation to the optimal detection scheme leads to a two-stage carrier synchronization architecture. The first stage produces tentative decisions and removes the modulation from the received signal. The second stage implements a smoothing filter for the resulting signal, providing an improved carrier estimate. Three basic first stage structures are discussed which provide different advantages. A first stage decision directed (DD) architecture provides the most spectrally efficient form of synchronization while still yielding good performance. A first stage PSAM structure provides the best performing technique without the spectral advantages of the DD structure. Finally, a hybrid architecture provides a compromise scheme with a tradeoff between improved performance and better spectral efficiency. A closer approximation to the optimal detection scheme leads to some near-optimal symbol-by-symbol detection schemes. By more closely approximating optimal detection, these schemes are more complex than the two-stage algorithms. The keys to reducing complexity in these algorithms are the use of decision feedback, finite observation windows and thresholding. Two specific algorithms, the DFE$(K,N)$ and RS-DFE$(K,N)$ algorithms, are developed and analyzed in detail. Some of the advantages of these techniques are that they require short decoding delays, they provide the statistics useful in soft decision decoding and they can easily be extended for use with diversity combining techniques.
Degree
Ph.D.
Advisors
Fitz, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.