Research On Radar Signal Sorting Technology

With the development of modern information countermeasure,the performance of radar reconnaissance system has higher requirements.As an important application in reconnaissance system,the main purpose of radar signal sorting is to de-interleave the received pulse sequence,and to identify the modulation mode and evaluate the threat level of the extracted radar signal feature information,which provides important information for combat strategy and combat plan.In the development process of signal sorting technology,radar system is also emerging,which leads to the variety of radar signals,the complexity of modulation mode,and the difficulty of radar signal sorting.Aiming at the radar signal sorting technology in complex electromagnetic environment,this thesis mainly studies the following aspects.Firstly,the parameters are extracted from the pulse data received by the reconnaissance receiver to obtain the pulse description word(PDW).On this basis,the inter-pulse modulation mode of radar signal is analyzed.For several common inter-pulse modulation modes,several traditional radar signal sorting algorithm models,such as dynamic extended correlation method,histogram method and PRI transform method,are also studied in this thesis.Due to the advantages of easy realization,stability and practicability,the sorting method based on dynamic clustering and sequence difference histogram(SDIF)is applied to practical engineering.The improved algorithm is encapsulated into software,which can achieve the purpose of improving the practical application value of the algorithm.Then,the dominant sets(DSets)algorithm is applied to complete the clustering work.In order to solve the problems existing in the current clustering algorithm,such as the performance of the algorithm depends heavily on tolerance parameter setting,threshold selection and prone to local optimal solution.In view of the shortcomings of DSets algorithm,the radar signal presorting algorithm based on DSets-DBSCAN is proposed in this thesis.DSets non-parametric clustering and density-based spatial clustering of applications with noise(DBSCAN)are combined to realize the non-parametric clustering of radar signal pulses.The simulation results show that the improved algorithm can complete automatic clustering without radar signal prior information,and achieve better sorting performance in the case of high false pulse ratio.Then,plane transform technology is introduced into the field of radar signal sorting to solve the problem that the existing radar signal sorting algorithm is sensitive to pulse loss.According to the characteristics of the image after plane transformation of the modulated radar signal,the diluted pulse based on the DSets-DBSCAN algorithm after pre-sorting is mainly sorted.In order to solve the problem of high time complexity caused by plane transformation sorting and traversal,an automatic searching method of transformation width based on plane entropy is proposed in this thesis.Based on the proposed method,image pre-processing is carried out on the transformed curve,and then the feature curve is extracted by density clustering algorithm.Data analysis and simulation experiments show that the proposed algorithm can also show good sorting performance when the pulse loss is serious.Finally,in order to construct a more perfect radar signal sorting system,this thesis combines the pre-sorting based on DSets-DBSCAN algorithm with the main sorting based on improved plane transformation,and proposes the idea and framework of the non-parametric clustering total score selection process.Through the simulation experiment,the proposed sorting method is compared with the conventional sorting method and the recently proposed sorting method,and it is proved that the proposed sorting method has better sorting performance when other conditions remain unchanged.At the same time,the measured data are input into the algorithm for engineering practice verification,indicating that the proposed method has practical application value.