Single Carrier Frequency Domain Equalization And Its Application In Short Wave Communications

Short wave communications is one of the most important communications technologiesin military domain. Due to the impact of transmission channels, signals at the receiver willsuffer from the inter-symbol interference (ISI) a lot, which may worsen the systemperformance. To reduce ISI, orthogonal frequency division multiplexing (OFDM) and singlecarrier time domain equalization (SC-TDE) are usually employed. But OFDM has a highpeak-to-average power ratio (PAPR) and it is sensitive to frequency offset and phase noise.SC-TDE also has a problem of computational complexity when there is a long time delayextension. In view of this, the thesis takes single carrier frequency domain equalization(SC-FDE) as the research topic.Firstly, the thesis introduces short wave communications and its wideband trend,analyses the short wave channel characters and two kinds of widely used modeling methods:Watterson model for narrowband and ITS model for wideband.Secondly, the thesis analyses the basic knowledge of SC-FDE, including its system blockdiagram and mathematical model. A commonly used frame structure in SC-FDE systems andthe unique word (UW) used as cyclic prefix are described. Types and generating methods ofUW sequences are discussed; the amplitude, auto-correlation and amplitude-frequencyresponse qualities are proved.Thirdly, theoretical analysis and comparison of PAPR, channel capacity, bit error rate(BER) performance and sensitivity to frequency offset and phase noise between SC-FDE andOFDM are done.At the latter, the thesis puts great emphasis on the frequency domain equalizationalgorithms in SC-FDE systems, derivations and simulations are made. Zero forcing (ZF)equalization and minimum mean square error (MMSE) equalization are two basic linearequalization algorithms. In addition, least square (LS) equalization is introduced; it has adisadvantage of training sequences requirement which makes the transmission efficiency lowand is not commonly used. To against channels with zero frequency response points, uniqueword aided linear equalization algorithms are proposed; by processing UW sequence and data sequence separately, the deterioration of performance can be efficiently relieved.To suppress the noise and ISI further, hybrid decision feedback equalization (H-DFE)was proposed. There is an equivalent algorithm which is called noise prediction decisionfeedback equalization (NP-DFE); it has independent feed forward part and feedback part.Residual inter-symbol interference cancellation (RISIC) algorithm is also a kind of decisionfeedback equalization, and it avoids the computation of matrix inversion.At the end, the thesis elaborates iterative block decision feedback equalization (IBDFE)algorithm in SC-FDE systems. It has a feed forward part and a feedback part both infrequency domain, by drawing iteration into equalization, system performance is significantlyimproved, but it has a high complexity. Although parameter approximation can reduce itscomplexity, but due to the parameter fuzziness, the performance is instable. So, the thesisproposes a new simplified algorithm using the MMSE criterion on the basis of IBDFEstructure, and at the same time, an equivalent form named noise prediction algorithm isderived too. Theoretical analysis and comparison on mean square error (MSE) of theproposed simplified algorithm and the existing low complexity algorithm and computationalcomplexity comparison of three kinds of IBDFE algorithms are made then. Results show thatthe proposed algorithm can reduce the computational complexity efficiently. Simulations onBER performance indicates that the proposed algorithm has a bit decrease when comparingwith traditional algorithm.