Study On Compressed Sensing Observation Method Of Underwater Target Active Acoustic Signal

Active acoustic signal to the lower echo processing is one of the commonly used methods,but the underwater environment is strange,all kinds of clutter(jamming signal)can be seen everywhere,and it is complex and changeable,these conditions give itself a certain error method in the underwater signal processing presents a challenge.In particular,the improvement of detection technology,the small target of high-speed motion,the detection accuracy of weak signal is more and more demanding,therefore,the treatment of underwater acoustic signal,weak target detection and processing is a difficult and hot issue in the field of signal processing.Under this background,this paper studies the compression sensing observation method of underwater target active acoustic signal,which provides the method and means to solve the problem of weak signal processing and large amount of sampled data,and then realizes the purpose of low speed observation and improved signal-to-noise ratio.The main research work is embodied in the following aspects:Firstly,based on the principle of underwater target active acoustic detection,the theory of compressed sensing is applied to the echo signal processing of underwater target,and the observation method of Gaussian matrix,Bernoulli matrix,sparse stochastic matrix,Toplize and cyclic matrix,partial hada matrix and so on is analyzed emphatically.The influence of sparsity,number of observations and so on on observation performance is realized and analyzed by simulation.Secondly,for the Gaussian matrix,Bernoulli matrix and other stochastic measurement matrices on the hardware system requirements,Toplize and cyclic matrix,part of the Hada matrix and other deterministic measurement matrix observation performance is not ideal,this paper is based on the rotational deterministic measurement matrix,by adjusting the coefficient of each column element of the measurement matrix to enhance the non-correlation between columns and columns,an observation method of generalized rotation measurement matrix is proposed,which is applied to the compressed sensing observation of underwater echo signal,and the performance of this method in weak signal processing and low speed observation is verified by simulating and analyzing the underwater echo signal with different Snr.Thirdly,in order to further improve the performance of the observation matrix in terms of construction time and signal reconstruction accuracy,this paper starts with the construction process of the known measurement matrix,and proposes a QR decomposition measurement matrix construction method based on cyclic direct product for the compressed sensing processing of underwater echo signal,which adopts the method of cyclic direct product from the finite one low-dimensional "The seed" vector constructs the high dimensional orthogonal matrix and obtains the orthogonal matrix through the QR decomposition,and finally selects the corresponding number of rows as the measurement matrix,and then improves the nonlinear correlation of the column vectors in the construction process.The simulation of underwater echo signal and the processing results of measured data show that this method has certain advantages in construction time,reconstruction precision and signal-to-noise ratio improvement.Finally,in order to realize the engineering application of the compressed sensing observation method in active acoustic detection,this paper takes into account the simplicity of physical realization,in order to reduce the hardware complexity of observation structure under the framework of compressed sensing,the compressed sensing method of direct default rate is adopted,and through the transformation of the original Nyquist acquisition system and the acquisition equipment based on direct underpayment is developed and the direct extraction sampling of low speed stochastic non-uniform is realized,and according to its sampling principle,the measurement matrix is constructed,the observation signal is reconstructed,and the process of signal acquisition,processing and reconstruction based on compression perceived observation method is realized.