Research On Target Detection Of Passive Radar Under The Condition Of Impure Reference Signal

Passive radar uses existing civilian or military signals as illuminators,through clutter cancellation,coherent integration and constant false alarm rate between reference signal and echo signal,target detection in region of interest is realized.Due to the characteristics of silent reception,separate transmitter and receiver,uncooperative signal,passive radar has broad application prospects in the fields of anti-jamming,anti-stealth,anti-low-altitude attack,anti-radiation missile,and spectrum sharing and utilization.In passive radar target detection,the ideal pure reference signal is the original transmission signal,however,due to the uncontrollable illuminators,reception deviation,target position and other factors,the reference signal will be polluted by other signals and become impure.This paper is engaged in the problem of the target detection under the condition that reference signal is not pure,and the main contents are summarized as follows:(1)Aiming at the problem that the reference channel is contaminated by target echoes,which leads to the decline of the reference signal purity and the performance of traditional target detection methods,a new received signal model is proposed.Firstly,making mathematical modeling for the reference signal,then the process of target detection which based on new signal model is theoretically deduced,in addition,the disadvantages of the new signal model to the traditional process of the target detection are analyzed,the target echoes involved in the reference channel will lead to arising a series of interference targets that are difficult to eliminate in the process of the detection,and will also reduce the detection ratio of signal to noise(SNR)of real target.Finally,the theoretical analysis results are verified by the simulations,and the results of the simulation show the correctness of the theoretical analysis.(2)Aiming at the problem of how to judge the impure condition of reference signal,a classification model based on neural network is proposed.Firstly,the problem of identifying whether the conditions of reference signal is pure or not is transformed into an image classification issue.Then,the data sets are built and the classification model is designed,trained and tested based on the automatic machine learning framework.At last,the performance of the classification model is verified by simulations.(3)Aiming at the problem of how to detect the target when the reference signal is not pure,a reference signal purification method is proposed,which first location and then reconstruction.Firstly,the impure signal involved in the reference signal is located by using the results of the traditional target detection.Then,the signal is reconstructed based on the idea of subspace projection,furthermore,filter out this signal by subtracting from the original reference signal.The results of the simulation show that the proposed method can effectively alleviate the adverse impact of the impure reference signal on the traditional target detection and improve the performance of the target detection.