An improved sparse underwater acoustic OFDM channel estimation method based on joint sparse model and exponential smoothing

Abstract

Compressed sensing (CS) based sparse channel estimation which sufficiently exploits the inherent sparsity of channels can achieve desirable performance compared with conventional channel estimation algorithms. However, in channel estimation common recovery algorithms, such as orthogonal matching pursuit (OMP), uses pilots inserted in one individual orthogonal frequency division multiplex (OFDM) symbol every time without considering the correlation of channels. Besides, estimated channels contain strong ambient sea noise. Therefore, accuracy of sparse channel estimation can be further improved through using both correlation and denoising of channels. In this paper, a channel estimation method which combines joint sparse model (JSM) and exponential smoothing (ES) is proposed for OFDM system. The proposed method mainly consists of two steps. Firstly, the UWA channel is estimated by joint sparse model based recovery algorithms, where the channel estimation is modeled as a problem of joint sparse recovery. Secondly, to denoise the estimated channels, the exponential smoothing is applied. Simulation results evaluate our method and show that the proposed scheme outperforms the methods of using pilots inserted in one individual OFDM symbol about 2 ∼ 3 dB.