An underwater acoustic frequency estimation algorithm based on first-order Newton iterations

Aiming at the problems of multipath interferences with different doppler frequency offset and the picket fence effect caused by fast Fourier transform, an effective underwater acoustic frequency estimation algorithm based on first-order Newton iterations is proposed in this paper. Bandpass filter is used to coarsely eliminate the multipath interferences with different doppler frequency offset, fast Fourier transform is used to roughly estimate frequency of the direct wave and the estimated frequency value is used as the initial value of first-order Newton iterations to accurately estimate frequency. The frequency estimation error caused by the picket fence effect is greatly eliminated through two iterations, achieving accurate frequency estimation of the direct wave. The good results of simulation (frequency of the transmitted single frequency signal is 10kHz, the sampling frequency is 50kHz and the transmission time is 0.08s) and velocity estimation experiment in a pool (the transmitted frequency was 10kHz, the sampling frequency was 48kHz, the transmission time was 0.1s, the initial horizontal distance and relative moving velocity between the transceiver nodes were 15 m and 1m/s, respectively) verify the effectiveness of the proposed algorithm.