Research On Radar Signal Sorting & Tracking Based On Data Mining Technology

Radar signal sorting and tracking is the core technology in electronic warfare.With the continuous development of electronic technology,radar technology and radar counter-counter measures technology,the modern electronic warfare signal environment presents the following characteristics:The number of radiators and signal density are increasing.New radiator systems are constantly emerging,and the waveforms are complex and changeable.The bandwidth of radiators is constantly increasing,and the overlap in the time domain is serious.In such a complex electronic warfare signal environment,the signals of multiple radiators overlap severely in various signal parameters.The traditional time-of-arrival-based signal sorting and tracking methods can no longer effectively extract the information in the radar signal data.This topic focuses on the signal sorting and tracking technology in the complex electronic warfare environment,combined with a number of technologies in the field of data mining to carry out research.First,the pulse description word and its composition used for signal sorting and tracking are explained.The signal environment is analyzed and explained from the perspectives of signal environment characteristics,signal change patterns and modeling of the signal environment.And the principle of the signal tracker based on pulse-repetition-interval is introduced.This topic focuses on the signal pre-sorting technology based on clustering.The commonly used fuzzy clustering pre-sorting algorithm is introduced.In practical applications,it is found that there are problems such as high algorithm complexity and manual setting of threshold parameters.In order to solve the problem of high complexity of the algorithm,combined with the disjoint sets and extracting clustering information with disjoint sets instead of tracking method,a low-complexity fuzzy clustering algorithm is proposed,and its time complexity is reduced to O(n2),and the space complexity is reduced to O(n).The simulation results show that the low-complexity fuzzy clustering algorithm is essentially same as the fuzzy clustering algorithm,and the complexity of the algorithm is greatly reduced.In order to further reduce the time complexity of clustering pre-sorting and at the same time enable the algorithm to automatically determine the threshold parameters,based on the low-complexity fuzzy clustering algorithm,the range search in the fuzzy clustering algorithm is changed to the knearest neighbor search in the λ neighborhood,and at the expense of space complexity in exchange for time complexity reduction.At the same time,based on the characteristics of the k distance graph,a threshold determination method based on the k distance graph is proposed to form a connected k nearest neighbor clustering algorithm.The time complexity of the connected k-nearest neighbor clustering algorithm is O(k·n·lbn),the space complexity is O(k·n),and it has the characteristics of low time complexity and automatic threshold determination.The simulation results show that the connected k-nearest neighbor clustering algorithm and the lowcomplexity fuzzy clustering algorithm that use the Calinski-Harabasz index to determine the optimal threshold are not far behind the sorting effect,while the time complexity is greatly reduced.This topic also analyzes the applicable scenarios of low-complexity fuzzy clustering algorithm and connected k-nearest neighbor clustering algorithm.In addition to clustering pre-sorting techniques,this topic also studied classification-based signal tracking techniques.Based on the C-SVC and One-Class SVM provided by libSVM,the SVM signal tracker is designed and implemented,and the selection of signal feature parameters involved in training and the determination of SVM classifier parameters are analyzed and tested.The simulation results show that the designed SVM signal tracker can effectively complete the signal tracking in the complex signal environment.Finally,based on the research content of this topic,a signal sorting function library is developed by C++,which includes:①The generation of the radiator(such as radar,decoy,noise)signal and the generation of the signal environment.②Clustering analysis algorithms such as low-complexity fuzzy clustering and connected k-nearest neighbor clustering.③Signal sorting algorithms such as SDIF and PRI Transform.④Signal tracking algorithms such as PRI signal tracker and SVM signal tracker.The simulation software of radiator signal sorting and tracking based on signal sorting function library was designed and developed by using Qt.This software simulates the parallelism of radiator signal emitting,signal tracking and signal sorting in actual electronic warfare,and realize the processing logic simulation of signal sorting processor.This software is used to simulate signal sorting and tracking in a complex electronic warfare environment.The results show that the algorithm proposed in this topic can effectively complete signal sorting and tracking,and the polarization information participating in signal sorting can effectively help distinguish radar and decoy.