Research On Channel Estimation In OFDM-based Cooperative Communication System

The size of the mobile terminals in MIMO-OFDM system can not meet the requirements ofMIMO antenna distance and the cooperative communication technology makes single-antennamobile terminals in the large space distance share each other’s antenna to send and receivesignals by a certain way, in order to create the virtual MIMO communication environment andget MIMO’s space gain. The channel estimation in the OFDM-based cooperative communicationsystem provides the important channel state information for the other modules (such as signaldetection, power distribution and node selection, etc.), and the accuracy of channel estimationdirectly determines the overall system performance. So the channel estimation is one of the keytechnologies required for researching. This thesis analyzes the cooperative communicationsystems, and makes the in-depth research in the OFDM-based cooperative communicationsystem. The work is briefly stated as follows:1. Research on channel estimation in Amplify-and-Forward (AF) cooperativecommunication system with single cooperative node in the frequency-selective fading channel.In this research filed, the disintegrated channel estimation algorithm based on Minimum MeanSquared Error (MMSE) has been proposed, but only suitable for the Gaussian fading channel. Toovercome this problem, this thesis proposes a MMSE disintegrated channel estimation suitablefor frequency-selective channel and its simplified form. This algorithm takes full account of thechannel correlation and noise impact, and improves the channel estimation accuracy andastringency greatly. The simulation results show that, compared with no convergence of theLeast Square (LS) and DFT (Discrete Fourier Transform) algorithms, this algorithm can get asuperior performance of convergence.2. Research on channel estimation in AF cooperative communication system with multiplecooperative nodes. Frequency division multiplexing pilot scheme in the AF system with multiplecooperative nodes will cost too many pilots. To solve this problem, a channel estimationalgorithm based on discrete superimposed pilot is proposed. This algorithm makes everycooperative node share the same pilots, reduces the system’s pilot overhead, and improves theeffective data rate of the system. Theoretical analysis and simulation results show that, thisalgorithm can ensure the accuracy of channel estimation, and reduce the system pilot cost at ahigh SNR environment.3. Research on channel estimation in Decode-and-Forward (DF) cooperative system withmultiple cooperative nodes. The channel estimation with frequency division multiplexing pilotsmay be followed by a "floor effect" caused by the frequency-domain interpolation error. Toovercome this problem, a time-domain channel estimation based on orthogonal superimposedtraining sequence is proposed for more accurate channel estimation. This algorithm utilizes thecommutative law of cyclic matrix to estimate the channel state information in the time domain,and avoids effectively the “floor effect” caused by frequency-domain interpolation. Theoreticalanalysis and simulation results show that, this algorithm has a better performance than LS andDFT algorithms with frequency division multiplexing pilots.