| The active sonar detection(ASD)system exposes itself when emitting sounds,leading to the positioning and destruction of itself or even the entire underwater platform.With the increasingly fierce competition in the ocean,the covertness of the ASD system is facing more and more severe challenges.The bionic camouflage covert detection method disguises active sonar signals as cetacean calls,and realizes covert detection through camouflage and deception,which has the advantages of simultaneously taking into account covertness and long-distance detection,and has received great attention.In this thesis,toothed whales with wide distribution,rich types of calls and high sound source level are used as the bionic objects,and the research is carried out from toothed whale call characteristics analysis,bionic signal design,range and speed measurement methods,etc.The main works can be summarized as follows:(1)To solve the problem of poor anti-noise interference and the difficulty of extracting harmonic envelopes of the traditional waveform envelope extraction algorithm based on Hilbert-Huang transform,a toothed whale call waveform envelope extraction algorithm based on the short-time Fourier transform and time-frequency two-dimensional time-frequency bandpass filtering is proposed.Taking the spectrogram contour,which is usually referred to as the “contour”,as the filtering baseline,a time-frequency two-dimensional parameter short-time variable parameter two-dimensional bandpass filtering method is designed,realizing effective extraction of the fundamental and harmonic spectrograms of the toothed whale calls,and the effective filtering of marine background and biological impulse noise.A proportional function mapping relationship between the spectrogram and the waveform envelope is constructed,and the proportional conversion from the spectrum peak to the envelope amplitude is realized by the amplitude recovery coefficient of the window function.The bionic signal synthesis results of six types of tonal sounds show that the proposed algorithm can effectively extract the waveform envelopes of various types of noised toothed whale calls with or without harmonic.(2)Aiming at the problem that model parameters are difficult to change due to the lack of closed-form expressions in the traditional bionic signal synthesis models,and the difficulty of imitating various cetacean calls with high similarity due to the low generality of traditional models,taking the tonal sounds with a high proportion in the cetacean calls as the imitation object,a general bionic signal synthesis model and method is proposed,which has closed-form expression and can imitate various types of tonal sounds with high similarity.According to the shapes and monotonicity characteristics of tonal sound contours,the Power Frequency Modulation Bionic(PFMB)model and Sinusoidal Frequency Modulation Bionic(SFMB)model are designed,realizing a quantitative expression of the nonlinear frequency modulation characteristics of tonal sounds.According to the characteristics of tonal sound contour with many extreme points and complex and changeable contour slope,a bionic signal artificial piecewise synthesis algorithm based on the splicing of multi-PFMB signals and multi-SFMB signals is proposed.Through tonal sound segmentation,PFMB or SFMB model matching and the splicing with continuous phase,the synthesis and imitation of various types of tonal sounds with high similarity are realized.Then,in order to realize the full automation of the synthesis of various types of tonal sounds,according to the short-time stationary and the long duration characteristics of tonal sounds,a bionic signal automated piecewise synthesis algorithm is proposed,which is realized based on automatic contour extraction and piecewise matching of PFMB model.Through the merging of connected areas of time-frequency points,the length threshold determination of each harmonic,the smoothing and segmentation of contours and the optimal matching of PFMB model,the automatic and high similarity imitation of various types of tonal sounds is realized.Finally,a bionic signal automated synthesis software based on C++ is designed.The experimental results show that the range and speed measurement performance of a bionic signal can be significantly improved by fine-tuning the expression parameters of the signal.Besides,the Pearson correlation coefficients between tonal sounds contours and bionic signal contours are higher than 0.95,which verify the camouflage of the bionic signals.(3)Focusing on the non-harmonic hyperbolic upsweep whistles,which have a high proportion in the false killer whale whistles,a hyperbolic frequency modulation(HFM)-like bionic sonar signal model with superimposed carrier frequency is constructed.A camouflage covert detection method is proposed,in which two sonar signal pulses are combined to measure distance and speed.Aiming at the limitation that the contour curvature cannot be adjusted under the fixed frequency range and duration of the traditional HFM signal model,a HFM-like bionic sonar signal model with adjustable contour curvature is constructed by superimposing the carrier frequency factor for adjusting the curvature.According to the advantages of high range resolution and large Doppler tolerance of HFM-like bionic sonar pulses,two sonar pulses are combined to be one pulse.In the measurement process,one sonar pulse with high range resolution is utilized to measure the target range,and two combined sonar pulses with large Doppler tolerance and Doppler iterated correction algorithm are utilized to measure the target speed.The experiment results show that when the signal to noise ratio(SNR)is-10 d B,the target range is 10 km and target speed is 10m/s,the root-mean-square errors(RMSEs)of the estimates of range and speed are 12 m and 0.032m/s,respectively.Moreover,the camouflage of the bionic sonar signals is verified by Support Vector Machine(SVM)classification method.(4)Aiming at the time interval hopping and center frequency hopping characteristics of sperm whale clicks,a camouflage covert detection method is proposed based on the mixed coding of Costas time-hopping and Costas frequency-hopping.By matching Costas time-hopping with the time interval hopping of the clicks,and matching Costas frequency-hopping with the center frequency hopping of the clicks,a bionic sonar signal model with mixed coding of Costas time-hopping and Costas frequency-hopping is constructed.Using Fourier Transform and inverse Fourier Transform,the Costas frequency-hopping coding is realized in the frequency domain based on the center frequency of the clicks,the Costas time-hopping coding is realized in the time domain based on the time interval of the clicks,thereby achieving the synthesis of the bionic sonar signals with high range resolution,high speed resolution,high mainlobe to sidelobe ratio and low Doppler tolerance.Based on the cross-ambiguity function and two-dimensional matched filtering,a single bionic sonar signal is used to achieve accurate measurement of the target speed and range.The experiment results show that when the SNR is-10 d B,the target distance is 10 km and target speed is 10m/s,the RMSEs of the estimates of range and speed are 91 m and 0.0086m/s,respectively.Moreover,the camouflage of the bionic sonar signals is verified by the neural network classification method. |
PDF Full Text Request