LSS-Target(the Low Attitude,Slow Speed And Small Target) Detection And Tracking Of LFMCW Radar

In recent years,with the popularity of various unmanned aerial vehicles,low-altitude security issues become increasingly prominent,which poses a serious threat to the normal life of the people and our national defense construction.In this severe situation,Linear Frequency Modulation Continuous Wave(LFMCW)radar has gradually assumed the responsibility of defending low-altitude safety with its excellent characteristics such as no blind zone,high resolution and low power consumption.And a series of radar target detection and tracking technologies have also become popular research directions in today's academic circles.Firstly,this thesis introduces the system structure of Linear Frequency Modulation Continuous Wave radar.The characteristics in time and frequency domain of the radar transmitted signal,received signal and intermediate frequency signal are analyzed.The basic principle of distance measurement of LFMCW radar is described,and the method of obtaining the target angle information by multiple receive antenna in LFMCW radar is introduced.Then the basic detection theory of Constant False Alarm Rate(CFAR)is studied.Various widely used CFAR detectors are introduced including mean level(ML)CFAR detector,order statistics(OS)CFAR detector and variability index(VI)CFAR detector.Due to the low altitude of LSS-Target,the working environment of radar is complex,while the detection performance of traditional CFAR detectors in complex environment is insufficient.Therefore,we propose a multi-segment ordered statistical CFAR detector and an improved VI-CFAR adaptive detector.Performance of these CFAR detectors is analyzed by Monte Carlo simulation in a variety of complex environments.And we detect the unmanned aerial vehicles in real environment.The effectiveness of CFAR detectors in low altitude complex environments is verified by experiments.Next,the advantages of Track Before Detect(TBD)for weak and slow-moving target detection and tracking are analyzed by comparing Detect Before Track(DBT)with Track Before Detect.This thesis selects TBD algorithm based on dynamic programming as the study object.TBD algorithms based on dynamic programming is simulated to detect and track objects with uniform linear motion in Gaussian background.In addition,the algorithm is applied to the radar hardware system to detect and track UAV in real environment.According to the test results in simulation and real environment,DP-TBD algorithm is improved and optimized combinded with the motion characteristics of UAV.The angle between the velocity direction of the current frame and the velocity direction of the previous frame participates in calculating the weights of the value function.We name this algorithm DP-TBD algorithms based on direction weighted.The algorithm is simulated and tested in a real environment.Compared with the unweighted traditional algorithm,the effect has been greatly improved.However,in some special cases,once the velocity of the previous frame deviates from the real direction of the target,basing on the direction of a single frame weighted is likely to make the track continue to deviate from the correct direction of target.In order to overcome the above-mentioned problem,we proposed an improved algorithm which is named DP-TBD algorithms based on multi-frame direction weighted.We use the angle between the velocity direction of current frame and the velocity vector sum direction of previous multi-frame to participate in calculating the weights of the value function,so that the Track Before Detect algorithms is more accurate and reliable.