Study On Related Algorithms Of Radar High-dimensional Perception For Non-cooperative Targets

With radar technology playing an increasingly important role in modern military, it is undoubtedly of great significance to further enhance the radar's perception ability so as to adapt to the increasingly complex battlefield environment and military targets. In current radar technology, the existing2D and3D imaging methods, such as SAR, ISAR and their3D imaging via multi-antenna interference, usually require the radar platform or target having specific movement patterns. Therefore, how to achieve the3D imaging for the non-cooperation movement of the radar platform and the target becomes a problem to be solved. In existing radar high-dimensional perception technology, to construct target's3D scattering center has become an important means for understanding the target comprehensive scattering properties and effectively realizing the target automatic identification. In existing scattering center3D imaging technologies, to achieve the3D geometric reconstruction of the scattering centers via associating the ID radar images under multi-aspect becomes the most simple and reliable measure and has been widely adopted. Presently, automatically associating the scattering center is the bottleneck problem that this technique has been faced, and the research of the3D geometric reconstruction for the unknown moving target's scattering centers is only in its infancy.Focusing on the problems of radar high-dimensional perception for non-cooperative target, the geometric analysis method is used to exploit the3D perception abilities of the single-antenna radar system under arbitrary non-cooperation movement patterns of the radar platform and target. In-depth researches are took place for the issues of non-cooperative target3D reconstruction based on scattering centers' range data, automatic association of the radar target scattering centers and the construction of the radar high-dimensional imaging system for non-cooperative target. The main contents and contributions of this thesis are as follows:Via the theoretical analysis for the algorithm model of the non-cooperative target3D reconstruction based on scattering centers'range data, a reconstruction model of "from Affine reconstruction to Euclidean reconstruction" is proposed and the factors affecting the reconstructed results are summarized. Meanwhile, two error models, i.e. Euclid additive error and Affine disturbance error, are built to inspect the reconstruction performance, and their quantization methods are proposed too. Consequently, the theoretical framework of the reconstruction algorithm is improved effectively.Focusing on the weak robustness problem of the existing3D reconstruction algorithms applied in complex backgrounds, several optimization reconstruction methods are proposed to improve the reconstruction performance. They include the optimization reconstruction algorithm based on motion path fitting,3D reconstruction algorithm based on radar's incomplete range data and optimization reconstruction algorithm for the condition of radar near-field. In these algorithms, with the optimization model for geometric reconstruction of the unknown moving target constructed, the bundle adjustment for the reconstructed parameters is successfully realized. Therefore, the robustness and applicability are improved effectively.Methods of radar target scattering center association are studied. Focusing on the conditions with known and unknown aspects, scattering center association algorithm based on geometric constraint mechanism are proposed which can be applied respectively to associate the scattering centers in the indoor controllable and outdoor unknown environments. Meanwhile, an automatic association system for non-cooperative targets in complex conditions is constructed. It can be applied for the non-cooperative target with unknown movement, and can accurately associate the1D scattering centers with missing, fault points existed under complex conditions. Thus, it has better applicability.A high-dimensional imaging system of non-cooperative target based on reconstructed path is constructed. With the ISAR signal model re-analyzed, the possibility of ISAR3D imaging for arbitrary3D moving target is verified, and a feasibility algorithm based on Radon-Wigner transform is proposed. Meanwhile, an imaging interval selection and motion estimation algorithm based on reconstructed path is proposed, by which the2D and3D imaging interval and motion parameters obtained can be effectively used in current ISAR2D imaging and3D imaging proposed in this thesis. With above methods applied, a high-dimensional imaging system based on reconstructed path is constructed, in which the scattering centers and unknown motion of non-cooperative targets can be perceived and the target2D and3D imaging can be realized with different motion patterns of the target selected and estimated.