Research On Ground Moving Target Tracking Technology Using An Airborne GMTI Radar

In the Gulf War 1991, the application of U.S. Air Force's JointSTARS achieved great success. The core of the JointSTARS is an airborne GMTI radar, which can not only quickly search and precisely detect ground moving targets, but also track multiple targets real time in a great area without being influenced by bad weather. Therefore, the GMTI radar attacks great notice, and now, the ground moving target tracking using a GMTI radar has became one of the most activity research areas in multi-target tracking. Considering the aerial target tracking algorithms which are the majority today have many differences from the ground moving target tracking algorithms, and also as part of the National Defense Advanced Research Project, this thesis addresses the data processing technology of ground moving target tracking using an airborne GMTI radar. The thesis focuses on nonlinear filtering, road constrained prior knowledge, VSIMM algorithm and ground formation target tracking. The main contributions of the thesis are as follow: 1. Considering the computational cost of UPF algorithm, which is an effective way dealing with the nonlinear filtering problems, the sqrt-root form of UPF algorithm is presented, and GMM is proposed in UPF algorithm. Simulation results show that these proposals can improve the performance of tracking while reducing the computational cost.2. A method for improving ground moving target tracking by using road constrained prior knowledge is presented. The problems of an on-road target tracking and a constrained nonlinear state estimation are consistent essentially, the unknown state vector of the target is constrained to a subset of the road segment vector space.3. A VSIMM algorithm is presented for ground moving target tracking. The road constrained model, the evasive Move-Stop-Move model and also the high maneuvering models are considered.4. The definition of formation target is given, and a new method based on affine transformation for ground formation target tracking is presented after the summarization of methods suitable for formation target tracking proposed by predecessors.