Study On Blind Equalization In The Communication Systems

Blind equalization technology is a newly developing adaptive equalizationtechnology. This technology can make the output sequence approach the inputsequence of the system by only using the transcendental information of thereceived sequence to equalize channel character without training sequence.In the fast developing information times, digital communication system andtechnology have become the important symbol of communication development.Digital communication system has been widely used in commercial applications.Consequently, corresponding digital signal processing techniques have beenwidely adopted in communication system, because these techniques can improveperformance and capacity of communication system.In order to meet high speed and great capacity information exchange, peopleput forward higher and higher demand. Intersymbol interference (ISI) is animportant factor, which affects quality and speed of digital communication. Thereason of ISI is the non-ideal characteristic of channel. Multi-path transmission isan important factor that can result in non-ideal characteristic of the channel.Adaptive equalization is a main solution of dealing with multi-path effect. It takesan important role in the process of information transmission. It can compensatethe non-ideal characteristic of channel and make high-speed communication thepossibility.In most digital communication systems that adopt equalizer, channelcharacteristic is sometimes unknown and time-varied. So the known trainingsequences are usually needed to be included in the data frames in transmitterwhen they are send to receiver. The reason for this is adjusting the coefficients ofequalizer to be fast constringed in wide range. However, the equalizer will inducesome problems. First, adding training sequences in the sending data will increasetransmitted spending, lower using rate of frequency band and affect efficiency ofcommunication system. More seriously, even the equalizer has been constringedand switched to working mode, after a few time, it's very possible to produceparoxysmal errors for the detector as the time-varied characteristic of channel. Soalmost all communication systems adopt the method of sending trainingsequences seasonally to retraining adaptive algorithm constantly, but this methodmore seriously lower the capacity of communication system. Secondly, in someapplications, it's impossible to provide training signal in transmitter such as inmilitary interception, earthquake deconvolution, image reconstruction andmulti-point communication network. For the reasons above, it's necessary tostudy a technique, which needn't known training sequences send in input andonly finish adaptive equalization based on output value of the system. This iscalled "blind equalization" technique. Similarly, the method that needn't trainingsequences is called "blind" method."Blind equalization" means estimating the unknown sending signals whenthere isn't any transcendent information about system and circumstance andneedn't training sequences. It is currently one of the key technologies in the fieldof digital communication, and is also a hot research issue in communicationengineering, signal and information processing, etc. Now, it has been used in thefiled of communication, radar, sonar, controlling engineering, earthquakereconnoitering, biomedicine engineering, etc, and there are important theoreticaland practical significance. Employing the technique of equalization to solve theproblem of ISI in high-speed digital communication system development hasimportant theoretical and practical significance, and it's surely a project withgreat prospect.The main work of the dissertation is to discuss the different technique ofblind equalization and carry out simulation analysis for them. Along withimproving some algorithms, the improved results are simulated and analyzed. Tosum up, the main conclusion and production as follows:1. The background of blind equalization is simply introduced in the paper,such as the procreant reason of ISI, the concept and basic of equalization. Allkinds of existing blind equalization techniques have been summarized. Bussgangblind equalization algorithm and blind equalization algorithm based onhigh-order cumulation quantity are emphases in the research.2. This paper systematically describes the basic theory of blind equalization,and sums up the blind equalization criterions as the zero-forcing criterion,kurtosis criterion and normalized criterion. The zero-forcing criterion can only beapplied theoretically, but cannot be put into practice. The kurtosis criterion has alittle practical value though it is constrained to some certain conditions. However,the normalized criterion is unrestricted and shows comprehensive potential ofapplication. The equalization criterion becomes a cluster rather than a single onebecause the order of cumulant can be chosen optionally.3. The basic knowledge of high-order statistic quantity is introduced in thepaper.4. This paper systematically analyzes CMA algorithm which is one of theBussgang algorithms and simulates the algorithm with Matlab. Aimed at thedefects of the traditional constant module blind equalization algorithm, this paperintroduces the design principles of the variable step size blind equalization, andderives the improved algorithm utilizing MSE as the step size control factor. Thenthe simulation results are provided to show higher convergence speed and smallerresidual error of the proposed algorithms compared with the traditional CMA.5. Analyzing the blind equalization algorithms of base on high-orderstatistics. According as normalized criterion, this paper applies the odd order(third-second order) normalized cumulant to blind equalization algorithms, anddeduces iterative expressions of algorithm. Giving a new method ofcounterchange based on the need of symmetric-to-asymmetrictransformation(SAT) in the transmitter and asymmetric-to-symmetrictransformation(AST) in the receiver while employing third-second ordernormalized cumulant algorithms to transport the symmetric signal. Then thesimulation results show that this algorithm can equalizes channel.Finally, we make conclusions for completed work and main contributions ofthis research work, and prospect for directions of future research.