Research About Tracking Fusion Algorithm Based On Particle Filter In Coordinated Detection System

Modern radar technology is becoming more and more powerful with new technology invented and new algorithm proposed. However, there still exist numbers of threats, such as low-altitude & high-speed target, stealth aircraft, radar jamming technology and anti-radiation weapons, which limit the effectiveness of radar, some of them even threatens the survival of radar. Coordinated detection system, which makes full use of diversity information like space, frequency and polarization, can achieve superior detection performance compared with traditional radar system, and owns the ability of adapting different scenarios, flexibility and survivability under adverse conditions. The key of coordinated detection system is information fusion of target tracking, and the algorithms of information fusion are quite different in different scenarios, such as linear array scene and wide-area distributed radar, etc.The weak target detection problem is addressed in this dissertation, and the working mechanism of coordinated detection system is studied, the fusion algorithms of target tracking in radar array scenario and distributed radar scenario have been focused on. The main content comprises:1. The concept of the coordinated detection system has been expounded, and the overall design of the system structure and the main research contents have been stated; the typical two scenarios, radar array and distributed radars, have been demonstrated, as well as the problems and characteristics in these scenarios.2. A particle filter algorithm based on the grating lobe gain is proposed to solve the grating lobe problem which occurs along array spacing increasing in uniform linear array. Other than the traditional grating lobe suppression method which brings broadening of main lobe, the algorithm, which is proposed from the perspective of signal processing, can fully develop the information of target carried by the grating lobe. Thus the algorithm get a better detection performance and more accurate tracking results.3. Distributed radars scenario is built, the advantages and disadvantages, as well as the applicable scope of different types distributed particle filter(DPF) algorithms are analyzed. A particle filter algorithm based on generalized cross covariance information fusion is proposed to address the asynchronous location problem of particles within distinct radar generated. The algorithm can obtain the near-optimal performance indicated by the centralized fusion algorithm; when the number of particles is unsaturated for the setting, the proposed algorithm can reduce the required communication bandwidth while ensuring the detection ability, furthermore, less number of particles imply less calculation cost of the system.The performance assessments of the above algorithms are carried out by the simulations. The results demonstrate that the proposed harness grating lobe algorithm and asynchronous particles fusion algorithm are effective and practical.