The Research Of Pilot-based Channel Estimation Technology For HF OFDM Systems

Channel estimation is one of the key technologies in HF OFDM system, the property of which is directly influence on the quality of the received signal, so the researching on the channel estimation technology has vital practical significance. This thesis mainly studies pilot-based channel estimation. Aiming at the issues in present algorithms, this thesis proposes an improved DFT-based pilot-symbol-aided channel estimation algorithm and an iterative interpolation algorithm. The main work of this thesis is summarized as follows:Firstly, an improved DFT-based pilot-symbol-aided channel estimation algorithm is designed based on the problem of high complexity LMMSE-based algorithm and the drawback of DFT-based pilot-symbol-aided channel estimation algorithm without completely filtering out noise. The improved algorithm utilizes principle of energy of channel response concentrated on a few large tap to introduce a threshold to select the significant signal while to filter out channel noise. Analysising wrong judgment of the situation when the threshold is introduced, To construct the error decision probability function related threshold, deducing an optimal threshold from the minimum total of error judgments probability. Simulation indicates that the channel estimation property of proposed DFT algorithm outperforms LS algorithm and traditional DFT algorithm,and its performance is close to the optimal LMMSE algorithm, the complexity of the algorithm is lower than LMMSE algorithm.Secondly, an iterative interpolation algorithm is proposed to compensate for the problem of the error of data subcarrier channel response which is far away from the pilot channel response in the linear interpolation algorithm. The algorithm utilizes the principle of the greater relevance with closer two subcarriers in the frequency domain response. First to divide estimated the subcarrier into two sets in order to first estimate channel response closer from the pilot subcarrier, then to estimate channel response far away from the pilot subcarrier. Then the idea of iteration is utilized to finish interpolation of the data subcarrier channel response. In iterative process, not only utilizing adjacent pilot subcarrier frequency response, but also using data subcarrier frequency response which is between with adjacent pilot subcarrier estimated. Simulation reveals that interpolation property of proposed iterative interpolation algorithm outperforms other interpolation algorithms under different pilot structure and different pilot interval. Proposed interpolation algorithm can fit the HF channel characteristic curve.