Research On Radar Signal Sorting Based On Channelization And DSP Implementation

In the increasingly complex electromagnetic environment,the problem of signal sorting has been paid more and more attention.As the signal density becomes denser,the types gradually increase and the parameters change faster,the traditional sorting method based on conventional parameters such as carrier frequency and pulse width are becoming more and more difficult.In the modern electronic warfare environment,the complexity of each parameter variation often leads to chaotic signal modulation,serious overlapping,and more and more serious threat.It is an indispensable part of electronic reconnaissance to correctly identify and sort all kinds of radar signals in this complex electromagnetic environment.In order to improve the sensitivity of the receiver,the channelization technology is first used for pretreatment before signal sorting.Aimed at the problem of high complexity of the narrow transition band channelized structure,put forward a kind of interpolation with NTB FRM unified modulation filter Banks(IFRM-UMFB)structure,the structure of two block filter interpolation for N times,further reduce the computational complexity of FRM filter group,and compared with classical FRM method,under the condition of the wideband digital receiver for eight channel reduced the number of 26.5% of the multiplier to use,to reduce the complexity.In order to improve the accuracy of sorting,the traditional clustering and sorting algorithms are optimized respectively.In order to solve the problem that it is difficult to grasp the number of cluster types and the selection of random cluster centers in complex electromagnetic environment,this paper proposes to cancel the pre-set cluster types and determine the initial cluster centers step by step according to Euclidean distance.Through experiments on several radar signals of different types,it is verified that their correctness and accuracy rate for each clustering category are all above 95%.In the stage of main sorting,aiming at the problem of limited processing ability of jitter signal in traditional sorting,the jitter analysis was carried out on the remaining data after SDIF sorting out sequence difference.The sliding window idea is adopted to reduce the probability of missing alarm caused by clutter.Finally,the two improved algorithms are combined with simulation analysis to test different types of radar signals,such as fixed,uneven and jitter,respectively,to verify the high accuracy and good performance of the sorting results.To make up for the defects of pulse analysis under complicated electromagnetic environment,improve the reliability of pulse analysis ability and the signal separation,the separation and identification in time domain and modulation domain multiparameter fusion,in the traditional separation to join the modulation domain parameters to separation,and gives examples to verify and validate the correctness of multiparameter sorting method and parameter modulation domain in order to achieve more subtle separation.Finally,on the hardware platform,k-means clustering and SDIF sorting algorithm are used to measure and verify various radar signals of different types.The experimental results show that the integrated sorting method and the implementation of hardware platform have good sorting performance and meet the requirements of the index.