Channel Estimation In MIMO-OFDM Systems

With not increasing the system bandwidth, the MIMO technology can extremelyenhance the system capacity and improve the using of frequency spectrum. OFDMis used as one kind of modulating technique which can effectively resist themultipath effect. The MIMO-OFDM technology which combines MIMO andOFDM technology has the merits of both of them. The MIMO-OFDM technology,one kind of new technique which uses array antenna to realize space-diversity inthe OFDM transmission system, can improve the signal quality. Using three kindsof diversity techniques which are time-diversity, frequency-diversity andspace-diversity, MIMO-OFDM can greatly increase the wireless system'sallowance to noise, disturbance and multipath effect.The wireless channels are not only frequency-selective but time-variant. Thechannel in the mobile communication system is usually a time-variant multipathfading channel. In MIMO-OFDM system, the merging-diversity, coherentdemodulation and decoding all need to know the parameters of the time-variantchannel. In order to obtain the higher system performance and correctly receive thedata from transmitters through the channel, the technology of channel estimation isessential to the MIMO-OFDM system. In this dissertation, the technology ofchannel estimation used by the MIMO-OFDM system is studied and analyzed.First, the coding technology of MIMO system is studied and the survey of MIMO-OFDMtechnology is introduced. Next, some characteristics of the wireless channel arediscussed, mainly focusing on the relations among the parameters offrequency-selective fading, time-selective fading and space-selective fading in thearea of small scale fading of the wireless channel. Then, in the single antennaOFDM system, based on three different pilot break-in modes, namely blockpilot-assisted channel estimation, comb pilot-assisted channel estimation andtime-frequency two-dimension pilot-assisted channel estimation, the comparisonsare made among the performances of different estimating algorithms. Finally, inthe basis of single antenna OFDM system, mainly for comb pilot andtwo-dimension pilot, the analysis is made about the technical performance ofchannel estimation of the MIMO-OFDM system.The main work and the conclusion of this dissertation are given as follows:1. Thoroughly study the dissemination characteristics of wireless fadingenvironment, and make a comprehensive understanding of the frequency, space andtime selective fading.2. Carry on comprehensive understanding to the space-time coding of the MIMOsystem.3. In the PSAM(Pilot Symbol Assisted Modulation)channel estimation ofOFDM systems, we derive different estimation algorithms and compare theperformances for three modes which are block pilot-assisted channel estimation,comb pilot-assisted channel estimation and time-frequency two-dimensionpilot-assisted channel estimation.The block pilot-assisted channel estimation method inserts pilot symbol into allsub-channels of one OFDM symbol according to the fixed OFDM symbol interval(usually one frame). It is only suitable for the instance that the channel changesslowly. In this pilot break-in mode, the channel is usually estimated with the LSestimation algorithm and the MMSE algorithm. But for the complex Gaussiannoise, the LS algorithm and the MMSE algorithm in time domain are equivalent.The comb pilot-assisted channel estimation inserts the beforehand known pilotsymbol in some sub-channels which positions are given, and in other sub-channels,the data information is transmitted simultaneously. It is real-time to reflect thechannel information in frequency domain. Under this pilot mode, the estimatingperformances are different with different interpolation algorithms. For the combpilot-assisted channel estimation, The four-points Bessel interpolation algorithm iswell applied in the environment of M.1225 pavements and in the static multipathchannel model. For two-dimension dissemination pilot-assisted channel estimation,with the interpolation first in frequency domain and then in time domain, thetwo-dimension estimation of the channel is finished. And when the linearinterpolation algorithm is used in time-frequency two-dimension situation, theestimation performance is better.It is two-dimension dissemination pilot-assisted channel estimation method thatinserting the known pilot symbol to the given points in time domain and frequencydomain. Firstly it estimates the channel parameter in the position where the pilotsymbol is inserted. Then it carries on the two-dimension channel estimation usingthe information parameter above and solves the problem by the way of thetwo-dimension interpolation filtering.4. This dissertation further studies the channel estimation method ofMIMO-OFDM systems combining with the STBC coding of Alamouti mode in thebasis of SISO-OFDM systems. The estimation algorithm derivation andperformance comparison are mainly based on comb pilot-assisted channelestimation and time-frequency two-dimension pilot-assisted channel estimation.5. One kind of optimized pilot structure is given, and the pilot's optimum designwork is finished by considering three-dimension information of space, time andfrequency.The outcome reaches the expectant effect by doing simulation tests for differentalgorithms.