Wideband HF-OFDM Channel Estimation Based On Compressive Sensing

The High-Frequency (HF) communication has been recognized as the primary means for long-range wireless communications since the advent of radio. With its advantages of far transmission distance and small system investment, HF communication plays an important role in military communication specially. However the HF channel is distorted by multipath delay spread and Doppler shift. In order to improve the transmission rate and oppose multipath delays, orthogonal frequency division multiplexing (OFDM) based on channel estimation and equalization is an effective technique, and channel estimation is particularly important.This paper gives the OFDM channel estimation algorithms for block and comb pilot based on Least Square (LS) and Minimum-Mean-Square-Error (MMSE), and this paper analyzes the advantages and disadvantages of the two different pilot structure. Besides it proposes a new interpolation algorithm based on multi-parallel to oppose the non-ideal synchronous.Recently, compressive sensing (CS) has been widely applied in many areas such as image processing, communication system and so on. With the continuous improvement of the transmission rate, we need to widen the HF channel band. As a result, multipath delays and Doppler become much more disgusting for channel estimation, which makes the LS channel estimation invalid. Hence, it is natural to consider using CS theory in pilot-assisted sparse channel estimation in solving this difficult problem also reducing the number of pilot symbols to raising the transmission rate further.At last, the non-ideal synchronous and the non-integer multipath delay will make the CS channel estimation inaccurate while we use the CS channel estimation in actual systems. So we propose an improved CS channel estimation based on noise-iteration-decision to oppose the influence of non-integer multipath delay, which makes the CS channel estimation widely used in the actual systems.