Research On Multi-frame Track Before Detect For The Netted Radar System

The netted radar system(NRS) carries out target detection and estimation based on cooperative work of radar nodes.NRS can greatly improve the detection and tracking performance by utilizing the transmitting and receiving resources of the system and the multi-direction scattered energy of the target.Usually,the detect before track(DBT) method is adopted to implement the NRS signal processing.That is,the single frame detection(SFD) is carried out on the echo data at each time,and then the detected points are used for subsequent processing.When the target signal-to-noise ratio(SNR) decreases or the target intensity fluctuates seriously,SFD will cause missing detection and lead to serious performance loss.In view of this problem,multi-frame track before detect(MF-TBD) is an effective method.MF-TBD jointly processes multiple frames of raw data at each time.It can retain the target information to the greatest extent,and improve SNR through the non-coherent accumulation of multiple frames.However,up to now,researches on MF-TBD are mainly focused on the single radar system.Its application to the multi-radar system faces a number of challenges,such as analysis of the theoretical performance,design of the data accumulation and fusion algorithm,and reduction of the calculation and transmission cost.Aiming at the above problems,this thesis studied the netted radar system MF-TBD(NRS-MF-TBD) method.The main work and contributions of this thesis are as follows:1.The closed form expressions of detection probability and false alarm probability of NRS-MF-TBD were derived.Based on the derived results,the theoretical performance prediction in typical scenarios can be analyzed.2.Two processing frameworks,the centralized processing and the distributed processing,were proposed.Under the centralized framework,the fusion center completes all data processing tasks,which can ensure the optimal performance.Under the distributed framework,radar nodes jointly process multiple frames locally,and then upload the processing results to the fusion center to complete information fusion.Therefore,it has higher computational efficiency and system flexibility.3.NRS-MF-TBD methods based on sliding window batch processing(SW-MF-TBD) and based on recursive processing(R-MF-TBD) were proposed.They can efficiently achieve the accumulation and fusion of the multi-radar multi-frame measurements,under the centralized processing.In addition,NRS-MF-TBD method based on the existence state estimation of the target(ESE-MF-TBD) was proposed.ESE-MF-TBD is able to suppress the noise interference,and finally improve the detection performance,when the target information provided by the multiple radar nodes is inconsistent.4.NRS-MF-TBD methods based on posterior information fusion(PIF-MF-TBD) and based on the plot sequence fusion(PSF-MF-TBD) were proposed respectively,which can effectively realize the information fusion of the multiple frames provided by the multiple radar nodes,and achieve the global target estimation.