Research Of Short-Wave Channel Estimation For OFDM System

Short-wave orthogonal frequency division multiplexing (OFDM) system can effectively resist frequency selective fading and improve the spectral efficiency. Channel estimation is of great significance to short-wave OFDM system and it is the basic of the equalization and demodulation. Because the channel condition of short-wave is bad and the spectrum is scare, high accuracy channel estimation with less number of pilots should be designed.The study of channel is the basic of designing channel estimation algorithm. In this paper, the characteristics of short-wave channel and the channel model are firstly studied. Then, we give the overview of OFDM channel estimations. And the basic principles of the wavelet de-noising and compressed sensing (CS) are introduced. At the same time, their applications in the channel estimation algorithms are introduced to make a fundament for the further study.The discrete Fourier transform (DFT) based channel estimation is suitable for short-wave communication because of its low complexity and high accuracy. But, the existence of virtual carriers in actual communication system will cause the leakage effect, which will enlarge the error and make the de-noising measures invalidate. In this paper, we first adopt an optimal suppression method to compensate the leakage effect and then make de-noising measures to further improve the performance. Comparing with the traditional de-noising algorithms, wavelet de-noising method can deal with the noise contained in useful signal and has better de-noising performance. So, we apply the wavelet de-noising method using the neighboring dual-tree complex wavelet coefficients to channel estimation algorithms. The proposed method combing wavelet de-noising and optimal suppression can effectively resist the leakage effect and reduce the noise. The simulation results show that comparing with other algorithms, the proposed algorithm has0.7to1.9dB gain at MSE and0.7to1.7dB gain at BER. And it can effectively resist the error floor and reduce the noise.The channel estimation using compressed sensing can estimate the channel information with less number of pilots, which has great practical significance for short-wave communication system. In CS based channel estimations, the adaptive sparsity matching pursuit (ASMP) based compressive channel estimation has bad anti-noise performance, although not need the information of sparsity. Because using the singular value decomposition (SVD) to modify the measurement matrix of CS can improve the robustness to noise. We use the SVD to modify the measurement matrix of ASMP based compressive channel estimation. The proposed channel estimation has better robustness to noise and low error. The simulation results show that comparing with ASMP based compressive channel estimation, the proposed algorithm has1.4dB gain at MSE and1.5dB gain at BER. And it has better robustness to noise and performance.The proposed two algorithms are suitable for short-wave communication. And our work will have some guidance and reference effect on the study of channel estimation for short-wave OFDM system.