Sparse Sampling And Multi-parameter Estimation Of Ultrasonic Signals Based On Piecewise Sinusoidal Coding Mechanism

Coded ultrasonic detection technology can improve the average sound power of the detection signal without increasing the peak emission voltage.This technology can also improve the signal-to-noise ratio with pulse compression technology.In recent years,it has been widely used in high sound-energy attenuation materials detection.In this dissertation,coded ultrasound is used for sensor array detection,and a piecewise sinusoidal coding mechanism is adopted for ultrasonic coding excitation.Based on the sparse sampling framework with finite number of degrees of information freedom,a Finite Rate of Innovation(FRI)sparse sampling method for piecewise sinusoidal coded ultrasonic signals is proposed.The transmitting array elements identification of echo signals is implemented while the amount of data collected is greatly reduced.From the two conventional parameters-echo delay and amplitude,it is added to the four parameters of delay,amplitude,signal discontinuity point and frequency and break through the multi-parameter estimation under the theory of sparse sampling of ultrasonic signals.Finally,the effectiveness of the proposed method is verified by simulation and actual test.The specific research work and conclusions of this dissertation are as follows:(1)The current research progress of coded ultrasonic technology and FRI sampling technology and their applications in the field of ultrasonic are analyzed.Combined with the characteristics and advantages of the two technologies,the general idea of FRI sampling and parameter estimation of encoded ultrasonic signals is established.(2)The common encoding mechanism and its characteristic is analyzed,put forward a kind of binary piecewise sinusoidal ultrasonic coding mechanism,established the mathematical model of the signal under this encoding mechanism,and the time and frequency domain characteristics and sensor response characteristic are analyzed.(3)The FRI sampling and parameter estimation methods of piecewise sinusoidal signals are studied,and the FRI sparse sampling and multi-parameter estimation of ultrasonic signals under this excitation mode are realized.The echo signal is divided into two channels.The first signal is directly modulated by exponentially regenerated kernel and FRI sparse sampling is performed.The discontinuous point of coded ultrasonic echo signal was estimated from the sparse sampling data,and the signal characteristic frequency estimation and source identification were realized indirectly by combining the coding mechanism.For the other signal,the high-order moment method is adopted to transform the coded echo signal into the ultrasonic pulse stream signal,and the signal is modulated by SOS(Sum of Sinc)sampling kernel,and then the FRI sparse sampling is performed.The echo delay and amplitude parameters are estimated from the sparse sampling data.The effectiveness of the proposed method is validated by simulation.(4)The ultrasonic detection platform of pipeline defects and the FRI sampling system of piecewise sinusoidal coded ultrasonic signal are established and the excitation encoding is based on a two-bit binary piecewise sinusoidal signal.From the sparse sampling data,the maximum estimation error of signal discontinuity point parameters is about 10%,and the frequency parameter maximum estimation error is16.8%,combined with the prior information of the encoding mechanism,the transmitting codes can be accurately identified.The pulse stream signals converted by high-order moment method greatly improve the range resolution and can effectively detect common pipe defects,the maximum estimation error of echo delay parameters estimated from sparse sampled data is 0.15%,and the maximum estimation error of echo amplitude is no more than 9%.The 2-bit binary piecewise sinusoidal encoding excitation signal takes 3.3MHz and 6.6MHz as the symbol frequency,and the Nyquist sampling frequency of echo signal is required to be no less than 13.2MHz.In the test,the two-channel sparse sampling rates were 7.5MHz and 1.2MHz,respectively.The sparse sampling data volume of one signal was reduced by more than 43%,and that of the other signal was reduced by more than 90%.