Research On Underwater Acoustic Doppler Factor And Time-Delay Estimation Based On Time-Frequency Analysis Of Multicomponent LFM Signal

With the development and utilization of marine resources and the rapid development of science and technology,underwater acoustic communication and detection technologies are receiving increasing attention,and its quality is seriously affected by the Doppler effect.Linear Frequency Modulation(LFM)signals are often used as sonar signals due to their better Doppler tolerance.At present,the Doppler factor estimation method for the single-component LFM signal at the transmitter side of the underwater acoustic channel has been deeply studied.With the widespread use of Multiple-Input Multiple-Output(MIMO)technology in terrestrial communications,MIMO technology has also started to be applied in the field of underwater acoustics,for example,using multicomponent LFM signals at the transmitter side of a underwater acoustics channel for underwater acoustic communication and detection.In this context,the existing single-component LFM signal Doppler estimation technology can no longer meet the needs of realistic underwater acoustic communication and detection,and it is necessary to study a Doppler factor estimation method suitable for multi-component LFM signals.To address the above issues,this paper proposes Doppler factor and time delay estimation algorithms based on the time-frequency analysis of multicomponent LFM signals,with the following main contents:1)The classical Doppler factor estimation methods and time-frequency analysis methods are analysed,and their shortcomings are summarised through simulation analysis of the currently popular Doppler factor estimation methods and time-frequency analysis methods,laying the foundation for subsequent research.2)To address the problem of broken residual time-frequency ridges in existing timefrequency methods when processing crossed multicomponent LFM signals in the timefrequency domain,this paper proposes an Incomplete Residual(INCRES)time-frequency analysis method.The INCRES method avoids the breakage of the residual time ridge by retaining some of the energy of the current estimated component signal when updating the residuals.Simulation results show that the INCRES method improves the accuracy of the timefrequency ridge estimation of crossed multicomponent LFM signals in the time-frequency domain.3)On the basis of the INCRES method,the(Incomplete Residual-Instantaneous Frequency Fragment)INCRES-IFFM time-frequency analysis method is further proposed.When the signal-to-noise ratio is low and the multi-component LFM signal with severe timefrequency crossover is processed,the time-frequency ridge estimated by the INCRES method is seriously distorted.To compensate for the limitations of the INCRES method,the fragmentary Instantaneous Frequency Fragment(IFFM)estimation method is proposed.The INCRES-IFFM method replaces the time-frequency ridge estimation method in the INCRES method with the IFFM method.The simulation results show that the INCRES-IFFM method outperforms the INCRES method in terms of time-frequency ridge estimation,FM slope estimation,and onset frequency estimation.4)Combining INCRES and INCRES-IFFM time-frequency methods,the Doppler factor and time delay estimation algorithm based on time-frequency analysis is proposed.Timefrequency processing of the received signal by the INCRES or INCRES-IFFM method to obtain the frequency modulation slope and starting frequency,and the combination of the two to obtain the Doppler factor and the time delay estimate.The simulation results show that the proposed method can satisfy both the Doppler factor estimation and the time delay estimation of singlecomponent and multi-component LFM signals.