Research On Passive Location Of Radiation Sources In Multi Sensor Formation Radio Stations

Passive positioning technology is now playing an increasingly important role in the military field.On the one hand,under modern information technology conditions,more covert target positioning methods are urgently needed for reconnaissance and detection?On the other hand,with the development of technology,reconnaissance and detection methods are gradually systematized and networked,which can more accurately obtain electromagnetic information and calculate the spatial position and motion status of targets.These also provide conditions for the development and application of passive positioning technology.This thesis mainly utilizes multiple receiving sensors to obtain Time Difference of Arrival(TDOA)and Frequency Difference of Arrival(FDOA)information for passive localization,and extends it to the localization and velocity measurement of multi target radio radiation sources in formation.The main content is as follows:1.For the estimation of time difference between the target and the receiving station under relatively static conditions,the basic correlation method has a decrease in estimation performance under low signal-to-noise ratio conditions.Therefore,this thesis studies various second-order statistical algorithms to improve signal-to-noise ratio,such as the generalized correlation method,quadratic correlation method,and Hilbert method,as well as high-order cumulant based time difference estimation algorithms,and compares their estimation performance through simulation in different Gaussian noise backgrounds.At the same time,in response to the disadvantage that correlation methods can only estimate integer multiples of time difference,fractional order estimation was studied using interpolation algorithms based on signal reconstruction and polynomial construction,and its accuracy was verified through simulation.2.For the estimation of time-frequency difference under the condition of relative movement between the target and the receiving station,if the correlation method is used to directly estimate the time-frequency difference,the result will be biased.Therefore,this thesis studies the joint estimation of time-frequency difference based on the maximum likelihood cross ambiguity function method.However,this algorithm has a huge computational load and poor real-time performance,and even by reducing the search range and other methods,it cannot effectively reduce the computational load of two-dimensional search.In response to this issue,this thesis studies a fractal dimension estimation method that converts two-dimensional time-frequency search into two one-dimensional correlations.The basic idea is to use the feature of the autocorrelation function being independent of time difference to extract frequency difference first,and then compensate a certain signal for time difference estimation.Finally,the accuracy and real-time performance of the fractal dimension estimation method under high signal-to-noise ratio conditions were verified through simulation.3.In response to the problem of data correlation in multi target time-frequency difference localization scenarios,this thesis studies an algorithm that utilizes TDOA to construct a cost function for preliminary localization.The basic idea is to grid the target area and construct a parameter matrix,then subtract it from the measurement matrix to construct an error matrix,and then search for local extremum points to obtain the initial position of the target.But when there are adjacent targets,this method cannot effectively identify them.In response to this issue,this thesis proposes a joint cost function method based on Direction of Arrival(DOA)and TDOA to improve resolution,and its effectiveness has been verified through simulation.By utilizing the matching of time frequency differences in the cross ambiguity function method,the velocity of the target can be estimated,and then Taylor expansion can further improve accuracy.The accuracy and stability of the algorithm’s positioning and speed measurement were verified through simulation comparison with classical algorithms.