|
The radar signal sorting is a key technology in the field of electronic warfare , and the signal sorting capacity to determine the radar reconnaissance system is able to adapt to the modern electronic warfare environment . Therefore, to improve the signal separation capacity of the reconnaissance system , real - time sorting and recognition , in particular in dense and complex signal the missing pulse rate larger case , is very important. This paper studies the radar signal sorting modern electronic warfare environment , first introduces the research background and significance of the radar signal sorting as well as domestic and foreign research and development and current situation . Then radar signal sorting parameters and their measurement accuracy , then the radar signal sorting key research the unknown radar signal sorting process , including radar signal pre - sorting and primary separation of two parts , pre- sorting algorithm using DOA, PW, RF combined sorting , dynamic clustering method is used , the specific algorithm : improved K- means clustering algorithm . The biggest advantages are: storage space without division , without setting the number of categories , which fits well with modern electronic warfare pulse density characteristics of an unknown number of unknown radar emitter . The main sorting is based on the separation of the PRI , the main research the correlation function method , the CDIF , SDIF , PRI transform method , modified PRI transform method , plane transform method . Nothing harmonic problems for the correlation function , CDIF and SDIF strategy to avoid the PRI transform method using suppression strategy , made ??two points correction PRI transform method : using a variable time the starting point , the use of overlapping PRI box . To adapt them to estimate jitter PRI , the simulation results show that the jitter 10% , pulse loss rate of 20% , it is estimated that the result is correct , then sequence retrieval and stagger to identify . Put forward a comprehensive program of separation , on the basis of the above work , program the following order : match with the deduction of the radiation sources are known , the remaining burst clustering for each class SDIF sorting not sorting , using the modified PRI transform pulse type identification , conventional , jitter , or mixed . Simulation results show that the integrated sorting program selected a very accurate measurement and sub ??parameters .
|