The Research Of Multi-parameter Radar Signal Sorting Algorithm In Complicated Environment

Radar signal deinterleaving technology is crucial in electronic warfare, and the level ofdeinterleaving is the measure of information processing technologies symbol ofsophistication of radar reconnaissance equipment. With the rapid development of radarinformation technology, the Multi-function and Multi-usage modern military radar withrandom agile technology is deployed a lot, causing the current electromagnetic environmentincreasingly dense, radar signal waveform change complicated and signal characteristicparameters space overlapped. The signal regularities utilized by the classical deinterleavingmethods were badly destroyed, making the performance of the traditional deinterleavingmethods reduced or even failed. These presented serious challenge to radar signaldeinterleaving.The traditional sorting methods hard-classification in either-or way cannot deal withthem effectively, ignoring the existence of overlapping samples intermediary. As a result, itcannot reflect the real world objectively. In response to these issues，a new radar signaldeinterleaving method was proposed which is based on cloud model joint geometry coveringalgorithm. Under taking full account of the radar signal vulnerable to noise pollution and theimpact of measurement error uncertainty case, this method can resolve the problem ofcomplex and ever-changing quickly as well as overlapping features serious recognition ofradar signal sorting. It can overcome the problem of the traditional neural networks and otherintelligent sorting algorithms over-fitting, local minima and difficult to determine the networkstructure, slow classification issues and so on. This sorting algorithm with the incrementallearning function can be more in line with the current electronic reconnaissance environment.In order to reduce the overlap probability of the radiation source parameter space, and tobetter describe the new system of radar signal, the inter-pulse signal sorting parameters wasan important research direction as the use of signal sorting parameters,So this dissertationpresents a radar signal deinterleaving method decision based on the optimal atom to changethe way of traditional time-frequency atom analysis methods, and it can lower the complexityof extracting classification features, as well as combine the cloud model and coveringalgorithm classifier. The simulation experiments are carried out for different types of radarsignals and similar signals of different parameters respectively, and good results are achieved. Finally, for the unknown overlap radar signals of radar intercept system were often facedwith, a new radar signal deinterleaving method based on data field and improved fuzzyclustering was proposed. Taking into account the overlapping signal intermediary, thismethod figures that the data field based on cloud model combined class of FCM method canchoose the initial center of signal samples to solve the initial value of the FCM algorithmeffectively and overcome the problem sensitively and easy to fall into local optimal solutionin traditional FCM method. Improving C means clustering algorithm based on cloud modelfor sorting emitter signal samples, better effect is achieved.