TBD Based Dim Target Detection Technology Of Airborne Millimeter Wave Radar

With the rapid development of the stealth technology and interference technology, detection and tracking systems are faced with targets in a low SNR. In traditional detect-before-track(DBT) methods, the threshold based decisions are used to the raw measurements. This will result in the loss of much information contained in the measurements. And the DBT methods can’t detect and track dim targets effectively. The track-before-detect(TBD) methods process a number of frames of the unthresholded data, and then declare the track and detection simultaneously. The TBD technology can take use of the time association of the data, and becomes an effective method to detect and track the dim targets. It has become the research hotspot in the field of detection and tracking. This paper studies the dynamic programming algorithm(DPA) and the particle filtering(PF) based TBD algorithms in the application to HPRF-PD radar and stepped-frequency radar.In chapter 2, the radar theory and the classical test theory are introduced. First, the extraction of target information and the detection range of the radar are analyzed. And then the theories of Bayesian detection, CFAR detection, and accumulation detection are introduced, providing theoretical support for subsequent chapters.In chapter 3, firstly, the superiority of DPA in detecting and tracking dim targets is introduced. And then the DPA-TBD algorithm is given in the application to HPRF-PD radar. The detecting and tracking performances of DPA-TBD algorithm and traditional accumulation method are compared through simulation experiment. Simulation results show that the DPA-TBD algorithm is superior to traditional accumulation method for targets of accelerated motion. And the DPA-TBD algorithm is not affected by target maneuver, while traditional accumulation method is very dependent on the track of the target. When the target maneuver greatens, the detection and tracking performance of the later method will degrade badly.In chapter 4, the dynamic model and measurement model based on HPRF-PD radar is established firstly. The theoretical derivation based on the recursive Bayesian estimation gives the probability density functions of the state and the existence probability. And then the implementation steps of the TBD algorithm based on particle filter are given and described in detail. Besides, the principle of using difference channel information for angle measurement is introduced. And the dual-channels PF-TBD algorithm is proposed. Simulation results shows that the DPA-TBD has lower computation cost and is easy for implementation, while the PF-TBD has a better tracking performance. And the detection performance of PF-TBD is better when the SNR is low. Also, the dual-channels PF-TBD algorithm can give the angle information of the target effectively. The detection and tracking performances of the two algorithms are almost the same under the given deviation angle. In addition the dual-channels PF- TBD algorithm has a shorter response time for the appearance of a target when the SNR of the difference channel is relatively high.In chapter 5, the imaging principle of stepped frequency radar is introduced firstly. And then the dynamic model and measuring model of extended target are established. By using a sliding window in the range profile and accumulating the energy within it, the energy diagram is obtained. After that the point target particle filter algorithm is used to detect and track the extend target in the energy diagram. The detection performances of PF-TBD algorithm and traditional threshold based method are compared through simulation experiment. Simulation results show that the detection performance of PF-TBD algorithm is significantly better than that of the traditional one. What’s more, the PF-TBD algorithm can track the distance, the speed and the energy of the target effectively.