Research On Denoising Of Acoustic Signal Based On Fractional Fourier Transform

The underwater acoustic channel has spatial-temporal-frequency dependent characteristic,which makes the development of underwater acoustic communication behind a lot.Research on underwater acoustic communication technology is becoming a hot topic for scientists and engineers in various countries.The result of underwater acoustic signal extraction will directly affect the development of underwater target detection,location,and tracking technologies.The instantaneous frequency of linear frequency modulation(LFM)signal changes linearly with time,which is used to be a carrier signal in underwater acoustic communication can improve the ability of the system which includes resisting the noise interference,resisting the multi-path interference and frequency selective attenuation.Fractional Fourier transform(FRJFT)is a very important algorithm in the field of digital signal processing in recent years.An LFM signal is a function on a certain order of the fractional Fourier transform domain.Hence,there is positive significance in introduction of FRET to underwater acoustic communication and developing a novel communication method based on FRET.The main research purpose of this paper is to using FRET to separate noises from LFM signals which has important applications in the field of underwater acoustic communication effectively.The main research contents of the paper are listed below:(1)Firstly,elaborate the definition and properties of fractional Fourier transform detailed.And compare and analyze three mainstream DFRFT algorithms.Then introduce the sampling DFRFT algorithm that proposed by Ozaktas is in this paper,and detailed mathematical derivation is carried out.(2)Combining the previous theoretical analysis and mathematical derivation,a hydroacoustic signal model was established.This paper discusses the LFM underwater acoustic signal filtering algorithm that based on FRFT detailed.The single-component LFM underwater acoustic signal and the multi-component LFM underwater acoustic signal are simulated and analyzed.The simulation results show that when the interference noise is strongly coupled with it,the filtering algorithm can separate noises from signal effectively.(3)Aimed at LFM signal mixed with Gaussian White Noise parameter estimation method in FRFT domain is a multi-extremal optimization process,which has the cont:radiction between calculation and estimation accuracy.To overcome this short-coming,a new method for parameter estimation of LFM signals based on improved particle swarm optimization and FRET is proposed.A two-level search method is used.Firstly,a rough estimation is made by directly searching with a larger step size.Then,using the estimated value as an initial value and the search area is divided into high-precision regions.The transformation order is considered as a particle,and using the PSO algorithm to search Optimal fitness function value,so as to estimate the corresponding parameter value of LFM signal.Simulation results validate that the new algorithm has fewer complexity and better estimation precision.(4)Aimed at the researches on FRFT mainly focus on the theoretical study.A fast fractional Fourier transform algorithm based on floating point digital signal processor(DSP)is provided.Based on the underwater acoustic communication system model and the proposed underwater acoustic signal model,an experimental system for anti-noise capability of underwater acoustic communication is designed.Relying on the existing resources of the laboratory,the validity and practicability of the fractional Fourier transform algorithm are verified on the hardware platform.By a large number of simulation and DSP hardware platform implementation to verify the effectively of the LFM signals achieve signal-to-noise separation based on the FRFT,which has high practical value of engineering.