Study Of Radar Target Track Initiation And Target Detection At Tracking Stage

The capability of fast attack of modern military weapons brings new threats and challenges for radar detection,and how to use limited system resource to improve target detection performance to the maximum extent has become the research focus of modern radar systems.This dissertation mainly studies the problem of target detection and tracking in monostatic radar system and netted radar system,and the purpose is to fully explore the detection potentiality of existing radar systems to obtain better performance of target detection and tracking,with emphasis on the idea of fine information processing and intelligent system design.The main research efforts are summarized as follows:1.A track initiation method in monostatic radar detection system is studied.During track initiation,Doppler information is considered to be used to decrease false track probability,which is reflected in the constraint of Doppler channel for stably flying targets and the radial range prediction method based on Doppler information,under the consideration of range ambiguity problem and Doppler ambiguity problem in Pulse Doppler radar system.Compared with the traditional track initiation method,the usage of Doppler information can effectively decrease false track probability,and hence improve track initiation performance.Besides,for the detection problem of weak targets,considering the contradiction that high false alarm rate can improve detection performance but will easily induce too many false tracks,a track quality evaluation method based on amplitude information is studied,and the mathematical evaluation models of plot quality and track quality are presented.By using the amplitude information and incorporating the session of track quality evaluation,not only the detection performance of weak targets can be improved,but also false tracks can be kept at a controllable level.2.On the basis of traditional monostatic radar detection system,target detection strategy at tracking stage is studied,and an integrated detection and tracking framework based on the concept of constant track false alarm rate is proposed,which uses target tracking information to adaptively adjust target detection strategy for improved detection and tracking performance.The core idea is to decrease detection thresholds in target's predicted region under the condition of constant track false alarm rate,and two methods of calculating detection thresholds are proposed,i.e.,the uniform method and non-uniform method.Numerical results indicate that,the proposed integrated detection and tracking method can successfully decrease detection thresholds in target's predicted region and highly improve target detection performance and track maintenance performance at tracking stage,on condition that target track can be quickly terminated when the target under tracking disappears,and finally a virtuous closed loop of detection and tracking is achieved for monostatic radar detection system.3.In multisite radar detection network at track-fusion level,for the problem of low track initiation performance in some radar sites with low signal-to-noise ratio,a cooperative track initiation method based on tracking information is studied.The core idea is to use the target tracking information provided by the radar site that has already accomplished track initiation,to decrease detection thresholds in target's predicted region under the criteria of constant probability of false track initiation.Numerical results show that,compared with the track initiation method without using tracking assistant information,the presented cooperative track initiation method can improve target detection performance and hence track initiation performance,and consequently enhance target localization accuracy of radar network.The presented method can not only be applied to multisite radar detection network at track-fusion level,but also be used for task handover between different radar sites,which shows high value in engineering applications.4.Based on the idea of integrated detection and tracking in monostatic radar system,a tracking-aided target detection method for multisite radar network at plot-fusion level is studied,where tracking information obtained in the fusion center is fed back to each radar site,to improve detection performance and tracking performance of radar network.Firstly asynchronous tracking method for radar network is studied,and detection strategy at local radar sites is given based on the feedback information from the fusion center,and then data association strategy in the fusion center is studied on the basis of the idea of multiple hypothesis tracking.Numerical results indicate that,by feeding back target tracking information in the fusion center to each radar site,detection thresholds in target's predicted region can be adaptively decreased at each radar site,and hence a better target detection probability can be achieved;meanwhile,each radar site can transmit target plots to the fusion center with a higher probability,and finally tracking performance in the fusion center can be improved,which realizes a virtuous closed loop of detection and tracking for the overall radar network.5.Based on the idea of integrated detection and tracking in monostatic radar system,a tracking-aided target detection method for multisite radar network at signal-fusion level is studied,where tracking information obtained in the fusion center is fed back to signal-fusion detector to improve signal-fusion detection performance and tracking performance of radar network.Firstly some frequently used signal-fusion detection methods are discussed,and the spatial registration problem which must be solved before signal-fusion detection is demonstrated,and a spatial registration method based on grid division is presented for the spatial registration problem at target tracking stage.Then tracking information obtained in the fusion center is fed back to signal-fusion detector to adjust target detection strategy,with improved signal-fusion detection performance and track maintenance performance,which realizes a virtuous closed loop of detection and tracking at signal-fusion level in the fusion center.