Research On Passive Time Difference Of Arrival Location Based On Second Correlation

Base on a lot of learning and studying about passive time difference of arrival(TDOA)location algorithm from many literatures,this paper introduces the basic knowledge,algorithm principle and calculation flow of the basic cross correlation(CC)algorithm and,it's improved algorithm,the second correlation(SC)algorithm for time delay estimation in detail.Then,the existing methods for the improvement of the SC algorithm are summarized and elaborated.After deeply researching the basic knowledge,properties and the algorithm itself of the SC algorithm,a method,named contrast second correlation(CSC),was proposed to solve the problem of lacking the quality evaluation method for time delay estimation,because a high signal-to-noise ratio(SNR)is needed when using the correlation functions.Based on the SC algorithm and the nature of the cross correlation function,this CSC method establishes a set of contrast second correlation group,and referred to as the contrast group.The original SC algorithm is referred to as the original group.Since the original group needs two signals and self-correlates one of the signals,in the contrast group,the signal which is not self-correlated in the original group is taken and self-correlated.In the process of subsequent delay estimation,by using the property of cross correlation function,adjust the calculation process to make the contrast group can get the same time delay value with the original group when there is no noise.Then,based on this CSC method,a further improvement is made,and a cross correlation operation is reduced after the improvement.In the case of actual noise,the noise between different monitoring equipment is relatively independent because of the distance between different monitoring equipment is very long,so it is considered that the noise carried by the signals received by different detection devices is independent.When the SNR is low,the quality of the data packet is evaluated by comparing the difference between the results of the original group and the contrast group.First,through the simulation,a preliminary selection of the judgment threshold of the contrast second correlation method is made,and the threshold is given.Then,the CSC method and the threshold are verified by measured data.The verification results show that the CSC method and the threshold proposed in this paper can meet the requirements of practical application.For the better data packets screened by the CSC method,when using the SC algorithm to calculation time delay,the problem of the multiple maximum peak points that may exist in the time delay estimation in measured data packet and the fence effect causing by fast Fourier transform(FFT).In this paper,the polynomial fitting is used to fit the peak points of the SC algorithm results and the accuracy of the calculated results can be pushed to the sub-sampling point level,so that a more accurate peak value and a higher time delay estimation precision can be obtained.When polynomial fitting is used to fit the peak point of the SC algorithm.It is difficult to fit because the peak results of SC algorithm is too large.To solve this problem,a method of linear normalization of fitting points before polynomial fitting has been proposed.Finally,the superiority of SC algorithm based on polynomial fitting is verified by measured data.The algorithm can solve the problems of multiple maximum peak points and fence effect,and can obtain more accurate time delay value compared with the SC algorithm under the condition of good data packet quality.