Research On Signal Sorting Algorithm In Passive Radar Seeker

Passive radar seeker is mainly used in anti-radiation missiles,and it needs to update the target's Angle information in a relatively complicated electromagnetic environment,so as to guide the anti-radiation missile to attack the target.Radar signal sorting is a process of detecting the target radar from the complicated electromagnetic environment and is an important part of passive radar seeker(PRS).This paper aims to improve the fast sorting ability of passive radar seeker for multiple types of coexisting signals.There are two parts in radar signal sorting,one part is the multi-parameter sorting,the other part is the single-parameter sorting.Traditional multi-parameter sorting mainly uses the pulse description words(PDW)to filter pulse and traditional single-parameter sorting just uses the time of arrival to calculate the difference and get the repetition period by setting appropriate threshold.In current complicated electromagnetic environment,there are many different forms of signals appeared at the same time,such as frequency agility,PRF staggering,PRF jittering,pulse width jittering,frequency jittering and so on.In the process of radar signal sorting,the multi-parameter sorting can case the accumulated error and the single-parameter sorting cannot adapt the frequency agility and the condition of a large pulse loss.For the multi-parameter,the content mainly focuses on the fuzzy cluster and support vector cluster(SVC).The details are as follows:1.To improve the robustness of single parameter sorting and make it adapt to various signal types,an improved algorithm based on harmonic to extract the PRI is proposed,which makes full use of the similarity of harmonic under the condition that the pulses are missing badly.According the characteristic of the various signal type,the result of sort is merged.2.To reduce calculation complexity in fuzzy cluster when the amount of data is large,the pulse extracted method is proposed.Part of the sample data other than all the sample data will be used to complete the process of fuzzy cluster.3.To reduce the calculation complexity in clustering labeling in SVC,the improved method based on SEP is proposed,which just makes the supports vectors as initial points to get the extreme points.Then direction vector is created to assign the rest points to corresponding extreme points.Compared with the SEP,the improved method reduces the average number of iteration steps for each data point to locate its corresponding extreme point via steepest descent process.