Research On Complex Radar Signal Sorting Method Based On Data Field And Time Domain Characteristics

The essence of radar signal sorting is to extract the pulse sequences with different characteristics from the large number of pulses intercepted by the reconnaissance system and to correspond them to each radar emitter.Correct sorting is the basis for the estimation and extraction of various radar parameters,which is necessary for the accurate target positioning,tracking and analyzing and further improvement of the data in the radar library.To begin with,this thesis introduces the basic characteristics of complex radar signals,and analyzes the time domain,frequency domain,space domain and amplitude characteristics of radar signal parameters so as to understand the working principle of the radar signals in essence.After that,the thesis further discusses the composition of the radar reconnaissance system as well as the composition,function and principle of the pre-sorting and main sorting machines of the radar signals.It is necessary for the traditional clustering algorithm to set the clustering center and its number in advance,and the sorting effect of complex radar signals is not so ideal.To deal with this issue,a radar signal sorting algorithm is proposed based on data field clustering.All the data samples are normalized at the beginning and the potential value of the sample is calculated according to the data field theory.The steepest gradient descent method is applied to find the maximum point and its number so as to determine the initial clustering center and its number,and then the clustering center is recalculated and iterated continuously until the convergence conditions are met.As for the main sorting of radar signals based on the time domain characteristics,the traditional histogram method and the improved CDIF and SDIF methods are first discussed and simulated,and the effects are analyzed.As the histogram method is less able to suppress the harmonics,the PRI transform method adds the phase factor to the autocorrelation function.In this way,the harmonics are suppressed,but this method is poorly adapted to the jittered PRI sequence.After the improvement,the PRI transform method can adapt to the jittered PRI sequence better by changing the starting point of time and overlapping the small boxes,which is testified by simulation experiments.In the end,a whole sorting process is presented,the sequence retrieved after the pre-sorting and the uneven recognized.The simulation results manifest that the algorithm is reliable with an ideal sorting effect.