Study Of Space-time Adaptive Processing Algorithms On Ground Moving Target Indication Radar

Up to now since Brennan proposed the concept of space-time adaptive processing (STAP) in 1973, numerous scientists, researchers and engineers have done a lot of works on STAP theory and its practical applications and significant progresses have been made as the rapid development of large scale integrated circuit technology, computer and digital signal processing technique. Nowadays, STAP has proven to be one of the best techniques capable of detecting weak moving targets in strong clutter environment and has been widely applied in ground moving target indication (GMTI) radar.However, because of the technical secrecy reason, the STAP details in real applications have seldom been introduced and discussed in published literatures, as well as complete GMTI simulation systems based on STAP algorithm. Spaceborne distributed GMTI radar can provide very large baseline and the minimum detectable velocity (MDV) for STAP algorithm can be much lower. Up to now the spaceborne STAP technology is relatively not mature and there are many challenges in its realization and applications. The performance of STAP is heavily deteriorated due to the undersampling in spatial frequency domain, so large grating-lobes and sidelobes are induced. On another hand, huge computation load is the major fact to prevent the development and application of STAP in real situations ever since the idea was proposed.The dissertation concentrates the study on STAP algorithms in view of the above problems and tries to find effective methods to solve them. The major work of the dissertation can be summarized as follows.Firstly, we discuss the basic idea of STAP algorithm and the principle, structure and performance of STAP processor. Common GMTI algorithms are compared with each other. Simulations show the outstanding performance of STAP algorithm. The impacts of several practical factors on the detection performance of STAP, e.g. jams, clutter subspace leakage (channel mismatch, dispersion, internal clutter motion(ICM)), yaw and nonlinear antenna array, are analyzed as well as the iceberg effect are discussed when these practical factors exist.Based on the principle of STAP algorithm, a GMTI simulation system is established for the uniform linear array (ULA) system. The clutter model, noise model and signal model of the simulation system are built, diagonal loading technique is discussed, as well as echo signal simulation and signal processing flowchart. Simulations are presented to show the effectiveness of the method and the reliability of the system.Then, we conduct study on the improvement and computational complexity reduction of STAP algorithm and the dimension-reduced STAP algorithm is discussed. Starting from the inverse problem of dimension-reduction, how to improve the performance without increasing the system dimension is discussed. We propose a STAP algorithm combined with APES technique to improve the performance of ordinary STAP algorithm. We apply APES method to two aspects of the STAP simulation system. First, we obtain the clutter characteristics by using APES approach. Then, APES approach is combined to STAP algorithm to improve the results of STAP algorithm. Simulations show that by applying APES the GMTI detection performance can be improved without increasing the number of antennas and pulses, at the same time no high computational complexity is involved.Finally, we study the reason why the blind-zone problem of STAP altorithm is introduced by spaceborne sparse array radar and applicable methods for solving the problem. On the analyzing basis of the general DPCA condition, nonuniform linear array with periodic structure destroyed and waveform diversity/frequency diversity techniques, we proposed a new STAP algorithm incoperating the idea of Minimum Redundancy Arrays (MARs) and Optimally Irreducible Arrays (OIAs), to solve the blind zone problem for space-based distributed aperture radars. Simulations show that the proposed aglrorithm can greatly reduce the number of blind zones and achieve better detection performance than that of waveform diversity algorithm with less widening of the notch of the SINR curve.