| Summary: | 碩士 === 輔仁大學 === 電機工程學系碩士班 === 102 === Multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM), intersymbol interference(ISI) is caused by a multipath fading channel. Cyclic prefix (CP) or zero padding (ZP) must be added to solve with this problem.
This research investigates the performance of the channel estimation when the virtual carrier (VC) is added in the CP-OFDM systems. In terms of methods used for channel estimation, subspace(ss) blind channel estimation methods have been widely applied. However, the restriction of this method is a large amount of OFDM symbols are required for channel estimation. The subspace channel estimation techniques was used to calculate the correlation matrix in the traditional systems. However, this matrix needs to process a huge amount of data for converging. In the study, a circular matrix was proposed to increase the number of OFDM symbols. We apply this method to the MIMO CP-OFDM systems; then design a equalizer of MIMO OFDM systems . The proposed method also applies to MIMO ZP-OFDM by using the overlap-and-add(OLA) technique. It is found that the proposed method can obtain a low BER close to the one with ideal channel, regardless of CP-OFDM or ZP-OFDM systems.
In the study, we propose three methods for channel estimation. The first two methods are traditional subspace channel estimation, and subspace channel estimation that is assisted by a cyclic repetition method(CRM). The last method, we combine subspace channel estimation and semi blind channel estimation.
Finally, the simulation results indicated semi blind channel estimation method is better at the low BER and increasing convergence rate than semi blind channel estimation method is better than cyclic repetition method(CRM).
|
|---|
