摘要

Acoustic signals propagating through shallow water usually present dispersive and multimodal characteristics. Traditional methods usually characterize the dispersion relationship in time-frequency domain or frequency-wavenumber domain based on arrays of hydrophones. Recently, some methods have been proposed to reduce cost with a single hydrophone configuration. In this context, we focus on the single-receiver case. To extract each mode from the received signal, the frequency-domain intrinsic component decomposition method is applied. Specifically, a parameterized model is built first to demodulate both the source phase and the modal phase in frequency domain. Then the amplitude is modeled as Fourier series and the coefficients are obtained by solving linear equations. With the estimated phase and amplitude, we can easily separate and reconstruct each mode. Since the source phase is also demodulated in our study, we can handle other sound sources (e.g., the chirp source), apart from the common impulse source. Both simulated and experimental examples are provided to verify the effectiveness of our method.

全文