The Mmse Channel Estimation Based On Dft In Ofdm System

In the fixture wireless communication field, how to provide a data service with high-speed, high-quality and variety is one of research hotspots. Orthogonal frequency division multiplexing (OFDM) has attracted a lot of attention because of its high data rate transmission capability, high spectrum efficiency and robustness against frequency-selective fading channels.Channel estimation is one of the key technologies for the OFDM system. An important advantage of the channel estimation is that it can make the coherent demodulation into possibility and improve the system performance to 3 dB. At the same time, the channel estimation is also used in the adaptive technology. According to the channel changes, the parameter of modulation mode, coding rate and transmission power can be changed, so as to transmit information farthest and improve the capacity of the system. Therefore, the channel estimation affects the system performance immediately and has a significant importance.At present, the channel estimation methods mainly include pilot-based channel estimation, blind channel estimation and pilots-aided channel estimation. In this thesis, only the pilot-based channel estimation method is investigated. Main channel estimation algorithms have Least Square (LS) algorithm, Minimum Mean Square Error (MMSE) algorithm and the Discrete Fourier Transform (DFT) based algorithm. The LS algorithm is relative easy and has a lower complexity. However, its mean square error is large due to the presence of noise and the Inter- Channel Interference. So, it is limited in the systems that need higher accuracy. The MMSE algorithm has well performance but the computational complexity is very high because of the matrix inversion. Meanwhile, it needs the channel statistical properties which are always unknown in practice. The DFT based channel estimation algorithm has a good tradeoff between estimation performance and algorithm complexity. It mainly uses the characteristics that signal energy in time domain is more concentrated than the frequency domain. The DFT algorithm has a better performance than the LS algorithm and a lower the computational complexity than the MMSE algorithm. Therefore, the DFT algorithm has attracted a lot of attention by researchers. In this thesis, the DFT based channel estimation is studied.An improved method based on the DFT algorithm is proposed. Though the MMSE algorithm has a better performance; it needs auto-correlation matrix of the channel and noise variance. In the present literatures, it is assumed that the channel statistical properties are known, which are unknown in the practical system. In this thesis, the noise variance is estimated in the time domain using the characteristics that signal energy in time domain is more concentrated than the frequency domain, and the channel autocorrelation matrix is estimated using the denoised channel impulse response. Then the noise variance and channel autocorrelation matrix can be used in the MMSE algorithm. Simulation results demonstrate the performance of the improved method is better than the DFT algorithm and closed to the optimal MMSE algorithm.