Parameter Estimation of Wideband Underwater Acoustic Multipath Channels based on Fractional Fourier Transform

• Published in: IEEE transactions on signal processing 2016-10, Vol.64 (20), p.5396-5408
• Main Authors: Zhao, Yanbo, Yu, Hua, Wei, Gang, Ji, Fei, Chen, Fangjiong
• Format: Article
• Language: English
• Subjects: Algorithms; Channel estimation; Channels; Delays; Doppler effect; Estimation; fractional Fourier transform; Mathematical models; Parameter estimation; Parameters; pulse compression; Signal processing; Underwater acoustics; Wideband; Wideband underwater acoustic channels
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2016.2582466

Because of relative motion between transmitter and receiver, the wideband underwater acoustic multipath channel can be more accurately described by a multi-scale multi-lag (MSML) model. The signal components received from different paths can be differentiated in terms of Doppler scales, time delays, and amplitudes. Estimation of these parameters is essential for many underwater applications. In this paper, by virtue of the fractional Fourier transform (FrFT), an accurate and efficient method for parameter estimation is proposed. The algorithm proceeds in an iterative manner, returning parameter estimates of the most dominant signal component successively. With the linear frequency modulation signal employed as the probe signal, the estimation of the Doppler scale factor of each dominant component can be converted into a process of searching for the optimal fractional angle/order of the received signal's FrFT. At each iteration, a sub-iteration is contained to adjust the optimal fractional order. The pulse compression techniques for LFM signal, in both time domain and FrFT domain, are utilized to obtain precise parameter estimates. Once the parameters of a multipath are estimated, the corresponding component will be separated from the received signal-through eliminating the resampled, delayed and attenuated version of the original probe signal. The sparsity of the underwater acoustic channel is considered to reduce the amount of calculation. Simulation results confrm that the proposed algorithm outperforms the existing FrFT domain filter and short-time FrFT domain filter in multicomponent separation, and that the proposed parameter estimation scheme surpasses the matching pursuit based method in accuracy.