Study On Equalization Technology In SC-FDE Systems

Single Carrier Frequency Domain Equalization(SC-FDE) system is an innovativecombination of single carrier technique and frequency domain equalization technique,which makes reliable communication with high transmission rate possible due to theelimination of inter-symbol interference(ISI) cased by multipath effect, it also avoidsthe complex operation in Single Carrier Time Domain Equalization(SC-TDE) systemsand the high Peak to Average Ratio(PAR) in Orthogonal Frequency DivisionMultiplexing (OFDM) systems. With its technical advantages, SC-FDE becomes theresearch hotspot in recent years, and it has been included by IEEE802.16(WiMAX)standards as one of the important physical layer solutions for the next generationcommunication systems.The thesis focuses on the frequency domain equalization technology in SC-FDEsystem, and discuss the methods which can achieve optimal balance between Channelestimation accuracy and computation complexity with the problems from the existingalgorithms in mind. The main tasks are displayed as follows:①Introduce the basic principle of SC-FDE system, summarize the keytechnology in frequency equalization including the equalization algorithm, the Channelestimation algorithm and the interpolation algorithm. Summarize the advantages anddisadvantages of the algorithms on estimation precision and computation complexity,which provide the entry point for the subsequent research.②Adjust the power of each part of the frame which is constant in traditional one,propose the method of power redistribution between UW block and FFT block in timedomain. Solve the issue of optimal distrubution by means of Energy distribution factordesigning. The simulation results show that, compare to the system before powerredistribution,the BER capacity is improved in the system after power redistribution inthe circumstances of same energy in a frame.③To solve the issues that traditional LS algorithm is not accurate and IFFT/FFTfiltering method is difficult to implement, we propose a improved Self-adaptionFiltering LS method based on the noise elimination in time domain. We employ aself-adaption filter, whose filter coefficient is designed by taking channel power andnoise variance into account to track the changes in channel, in time domain during thetransform domain interpolation in traditional LS method. Through the analysis of simulation results in the multipath channel, we find that Self-adaption Filtering LSmethod can obtain Higher channel estimation precision than the other two method.④Because the need for noise variance in MMSE equalizaion method, we proposea SNR estimation MMSE method based on the feedback structure with considering thefeatures of frame structure in SC-FDE system. The method employ ML algorithm toestimate the noise variance from redundant UW sequence in FFT block which has beenequalized using ZF equalization method. Again, use MMSE equalization method toequalize the FFT block with the noise variance obtaind before. The simulation resultsshow that, the proposed method can achieve the approximate BER capacity compare totraditional MMSE method without additional cost.