Data Fusion Tracking And Work Model Identification Of Radar

With the development of the electromagnetic battlefield-“The Fifth Battlefield”,the role of electronic warfare in the modern battlefield environment is becoming more and more important.In order to implement electronic warfare more purposeful and better,is an essential task that use radar to detect and track enemy target in the early stage more effectively,and in order to meet the growing operational needs,traditional single-source data tracking has been replaced by multi-source data fusion tracking which is more effective gradually.In addition,electronic countermeasures as a major component of electronic warfare,the identification of radiation sources and other issues which involved in electronic reconnaissance field in electronic countermeasures are also hot topics in current electronic warfare,and this play a vital role in accurate interference and defense in subsequent operations.Therefore,this paper takes electronic warfare as the main body,for the target tracking problem which before electronic warfare,we studied multiple heterogeneous sensors data fusion tracking.For the electronic reconnaissance part of electronic warfare,for some specific scene conditions,we studied radar emitter identification and radar working mode identification.First of all,we proposes multi-scale based single-target three heterogeneous sensors data fusion tracking method.For the tracking data of three heterogeneous sensors for laser,infrared and millimeter-wave radars that are available in some cases,based on the multi-scale idea,a multi-scale model is established according to the difference of sampling frequency of multi-sensor,and filtering,centralized and distributed fusion are sequentially performed to complete effective tracking of single targets.Secondly,we proposes the method of radar emitter identification based on SSAE and Light GBM and the method of radar working mode identification based on CFSFDP.For the complex and unknown radar pulse signal environment formed by multiple radar working mode pulse signals corresponding to various radar emitter that may be obtained in the battlefield,using the proposed Feature extraction and classification framework based on SSAE and Light GBM extracting and classifying depth features of radar pulse signal fuzzy modulation features firstly,the initial identification of the radar emitter is well done.Then based on the identification result of the radar emitter in the previous step,the improved CFSFDP method is used to perform secondary clustering on different radar working modes of the same radar emitter to achieve identification of the radar working mode effectively.Finally,we proposes the method of sample data enhancement of radar working mode based on GANs and the method of radar working mode identification based on neural network fuzzy decision fusion.Based on the modeling of the radar pulse group sequence,firstly,the GANs are used to enhance the unbalanced sample data for the sample data of the radar working mode which may be obtained in the battlefield.Then,based on the parameter overlap caused by environmental noise error in complex electromagnetic environment,the MRNN-FDF identification framework is proposed,which takes into account the overall and time series characteristics of the radar pulse group sequence,uses MLP and RNN to identify,and then blurs the decision results and use D-S evidence theory to fuse thus give the final decision,and complete the identification of the radar working mode effectively.