In order to estimate ship velocity, we have applied the multiple signal
classification (MUSIC) method to one-pulse burst-echo Doppler sonar data. The
MUSIC method enabled us to estimate the Doppler frequency shift precisely under a
low-signal-to-noise ratio (SNR) situation even from a one-pulse burst-echo signal with
small data points. In simulation experiments, a signal frequency component of f = 16 Hz
could be extracted from N = 128 data-point data under a 40% Gaussian-distributed
additive noise with a sampling frequency fs = 2048 Hz. From actual one-pulse
burst-echo signal data of N = 128 points, a Doppler frequency shift of Δf = 1.45 kHz,
corresponding to a ship velocity 1.76 knots, was clearly detected, the frequency resolution
of which was almost impossible to attain by the conventional Fourier transform (FFT)
method. We found that the MUSIC method was useful especially for estimating the ship
velocity at a very low speed.