Study On Synchronization Algorithm In OFDM System

First adopted in DAB, OFDM technology has been widely studied and used in DVB,WLAN and ADSL systems in recent years. And OFDM technology and its application indifferent areas have been regarded as more and more important. The amount of appliedfields using OFDM techniques is growing rapidly. But the problem still exists that OFDMsystem can be easily disturbed by the offset of carrier-frequency that leads to disturbancewithin sub-channels, thus destroys orthogonity among carriers. Therefore, how to obtainaccurate synchronization is the key problem for OFDM systems.Many sub-carrier signals adding to each other produces OFDM symbol, which differsone from other by orthogonity。Therefore it is important thing to assure orthogonity forOFDM system, which needs accurate carrier synchronization. There are three demands insynchronization: carrier synchronization that guarantees the equality of oscillationfrequency of the receiver to that of the sender;signal synchronization that keeps the samemoment for the beginning and ending of IFFT and FFT;sampling synchronization thatsustain same sampling frequency for both receiver and sender. Since samplingsynchronization cannot influence OFDF system seriously, only carrier synchronization andsignal synchronization are mainly discussed in this paper.There are two kinds of synchronization algorithm regarding to timing and frequencyoffset in OFDM system: one is non-blind estimation, based on pilot symbol, which canlead to the waste of pilot width and power resource, thus decreases efficiency of the system;the other is blind estimation utilizing structure of signal such as cyclic prefix and virtualsub-carrier, which overcomes disadvantages of resources waste. In this paper, a blindestimation algorithm based on cyclic prefix and a non-blind estimation algorithm utilizingtraining symbols are presented.1,An improved synchronization iterative ML algorithm based on cyclic prefixSignal timing synchronization and carrier synchronization are the key problems intelecommunication system. Especially, multi-carrier system is more sensitive on signalnon-synchronization and carrier frequency offset than single-carrier system. Almost all themulti-carrier system inserts cyclic prefix in order to eliminate ISI. Based on ML estimationalgorithm and using information within cyclic prefix, which can perform synchronizationand carrier frequency synchronization, an iterative synchronization algorithm is presentedin this paper. This algorithm does not use pilot and can overcome the disadvantages ofresources waste.A traditional algorithm in blind estimation algorithm is ML algorithm. ML algorithmhas several advantages such as small amount of calculation, low redundancy, easy to beimplemented. It can also estimate timing and frequency offset simultaneously. However,this algorithm can hardly be utilized in complex OFDM in application, since it has anarrow scope of frequency offset estimation in AWGN and is coarse in timing estimation.Therefore, the task of this paper is to sufficiently modify this algorithm through exerting itsadvantages and overcoming its disadvantages. For the reason mentioned above, combiningML algorithm and iterative algorithm together, an iterative synchronization algorithm isobtained with less calculation burden and easy hardware realization.Iterative ML algorithm seeks maximum by doing relative calculation using cyclicprefix with a better performance in AWGN. Nevertheless, it has its own disadvantagesunder the circumstance of multi-path fading channel. That is a portion of sample points ofcyclic prefix can be disturbed by ISI, which changes some relative value and influences theveracity extent of the algorithm as a whole. Therefore, the precision of iterative MLalgorithm has a long way to go with fading channel. In order to overcome the thisshortcoming, suitable sample points mending using relative calculation can be done forimproving the precision of algorithm and endowing the modified iterative ML algorithm abetter performance. An infection caused by channel to relative values is consideredSimultaneously to lessen the variety of amplitude of the sample points and make theiterative ML algorithm more accurate.In this paper, an improved synchronization iterative ML algorithm based on cyclicprefix is proposed. This algorithm considers the infection of multi-path fading channel andtakes advantage of the sample points having same position in many OFDM sample pointsadded with each other and amends the sample points used to compute the correlation.Therefore it can reduce the variety of amplitude of the sample points and improveprecision of the proposed synchronization iterative algorithm. This algorithm alsoimproves the weakness of the primary iterative ML algorithm in the estimation of timingand frequency offset in Additive Gauss White Noise (AWGN) channel and establishesaccurate synchronization in multi-path fading channel. Simulation was conducted in bothAWGN channel and multi-path fading channel respectively. Simulation results showed thatthe improved iterative algorithm has the better synchronization performance for bothchannels.2,A coarse synchronous method based on IEEE802.11aIn order to utilize data to obtain more accurate synchronization sufficiently in OFDMsystem, the synchronization process is divided into two segments: seizing and tracing. Thedividing makes it easy that only inter-segment tasks be considered without consideringcounterparts of other segments, therefore designing the synchronization structure can bemore liberate. Generally, in order to obtain better synchronization performance, seizingsegment is divided into two modes further, named coarse synchronization and delicatesynchronization. Coarse synchronization seizes in a large scope so that it can mainly assurethe precision of orientation, leaving the remained exiguous offset adjusted by delicatesynchronization.A standard IEEE802.a WLAN based on OFDM, which is used as physical layertechnology, has been widely concerned recently. It provides a data velocity of 56Mbps.This is a telecommunication standard, which applies OFDM in grouping subject. In thissystem, timing synchronization, frequency offset estimation and channel estimation arefinished by the two training symbols, which makes up of two parts: 10 short trainingsignals for coarse synchronization and 2 long training symbols for delicate synchronizationand channel estimation.Since there is still some sensitive problems in IEEE802.11a standard towards timingoffset and frequency offset, a coarse synchronization algorithm is presented as anothermain point in this paper, which utilizes training symbols similar with the one inIEEE802.11a. The algorithm can obtain more accurate timing synchronization anddecrease frequency offset to a narrower scope. It has such advantages as low complexityand high practicality. Simulation result shows the good performance of this coarsesynchronization algorithm.In order to obtain a method of timing synchronization and carrier synchronizationused in WLAN with a low complexity and high practicality, the proposed coarsesynchronization algorithm based on IEEE802.11a system is towards IEEE802.11a, itsadvantages can be applied in more fields through constant effort.