FRFT-based Parameter Estimation of Time-varying Wideband Underwater Acoustic Multipath Channels

The time-varying property and wideband nature are considered as the most difficult challenges of the underwater acoustic channels. Due to the low propagation speed of the sound, the relative motion between the transmitter and receiver can result in serious Doppler effect. Even if both of the platforms stay still, the multipath signals along different eigenrays vary in angles of arrival, hence the radial velocities. Such Doppler velocities brings different time compressions or dilations to the multipath signals. In addition, as the signal bandwidth is usually comparable to the center frequency, it is more appropriate for the underwater acoustic channels to be modeled as wideband channels. In this paper, the parameter estimation of the wideband time-varying underwater acoustic channels is investigated. Linear frequency modulation signals are used as the training sequences, and the fractional Fourier transform is the key of the estimation algorithm. The communication scenario, with a fixed transmitter and a moving receiver, is built by the Bellhop software. The generated channel impulse response shows obvious time-varying property, as well as sparsity, in the multipath amplitudes and delays. The estimation result recovers the channel impulse response very well, thus confirms the effectiveness of the proposed method.