Research On Channel Estimation For OFDM Cooperative Communication Systems

Cooperative communication technology can obtain spatial diversity gain without increasing the hardware devices. Orthogonal frequency division multiplexing(OFDM) technology is an effective method to fight against the effect of frequency selective fading channel. Combining cooperative communication technology and OFDM technology can take full advantage of both and further improve system performance. In OFDM cooperative communication system, channel estimation is the premise for coherent detection, link adaptation, wireless resource scheduling and allocation, etc. Channel estimation is also crucial to obtain diversity gain and enhance the data transmission performance in cooperative communication system. In this dissertation, the problem of channel estimation in OFDM cooperative relay system is considered and the research mainly focuses on how to estimate a cascade channel established by two terminal nodes via a relay node. Our main contributions in this dissertation are described in the following.Firstly, for OFDM cooperative communication system, three channel estimation methods are studied, namely, the least square, the linear mean square error and discrete Fourier transform. Their advantages and disadvantages are compared and analyzed. The results show that the DFT channel estimation is a compromise method between LS and LMMSE.Secondly, to solve the insufficient ideal noise filtering in the traditional channel estimation algorithm, an improved channel estimation algorithm with threshold noise filtering is proposed. Although the LS estimator is simple and easy to be implemented, its performance is relatively poor without any noise suppression. The traditional DFT channel estimation algorithm only eliminates the noise beyond the guard interval(GI) length, but not suppresses the noise within the GI length. The traditional threshold filtering channel estimation method can eliminate some noise within the GI length, which can play a good role in traditional point-to-point system, but it does not work well in the cooperative communication system. For this reason, the traditional threshold is improved. On the basis of the traditional threshold, the maximum point beyond the GI length is introduced, and let the mean of the above two values as the improved threshold. The channel estimation performance is greatly enhanced with the improved threshold in cooperative communication system.Finally, to further improve the estimation performance in cooperative communication, a novel DFT channel estimation method combing improved threshold with Mahalanobis distance decision analysis is proposed. Channel impulse response is considered to be two parts: CIR part and noise-only existing part, respectively. Moreover, the channel estimation is regarded as the binary decision analysis on channel taps and noise components. Thus, using Mahalanobis distance decision analysis can also enhance the estimation performance by effectively suppressing the noise within the GI length. The proposed algorithm first obtain preliminary effective channel taps by setting the improved threshold and then modify the detected channel taps by employing the Mahalanobis distance decision analysis. Results show that the proposed scheme is more effective than other existing schemes in terms of estimation performance.