Space-Time Adaptive Processing For Clutter Mitigation Of Space Based Radar

It is inevitable for the future battlefield target surveillance radar to utilize space based platform. However, due to the large coverage area of space based radar (SBR), the echoes from moving targets of interest could be totally buried in the strong clutter returns reflected from the earth. Therefore, to detect targets of interest, the strong ground clutter must be mitigated effectively. Clutter mitigation technique is a key technique for SBR. The research on this subject is of significant theoretical and practical valuese. This dissertation focuses on the space-time adaptive processing (STAP) method for clutter mitigation of space based radar.Firstly, a space based airborne moving target indication (AMTI) radar model is built by compromising various mutually restricted factors. The methods of modeling and simulation of space-time two-dimensional clutter for SBR are proposed in detail. According to a through analysis on the clutter mitigation performance of STAP methods for the space-based AMTI radar, two research emphases are pointed out, which are reduced-dimensional (RD) / reduced-rank (RR) STAP methods and the impact of the earth's rotation. Moreover, the partition of subarrays and the design of their weights are also very important when realizing space-time adaptive processors in practice.The local degree of freedom (LDOF) of clutter is a good reference for the choice of RD-STAP methods and the determination of the system DOF. In this dissertation, the LDOF formulas for four commonly used RD-STAP methods under ideal linear array and unformed subarray configurations are presented by rigorous proof. The influences of various operations and non-ideal factors in RD process on LDOF are analyzed. Further, the LDOF theorem is generalized to the space-based AMTI radar configuration and the clutter mitigation performance of RD-STAP methods for SBR is studied.The performances of RD/RR-STAP methods are confined by their arithmetical structures and the statistical clutter information they use. By introducing the preconditioned method in computational mathematics into the STAP filed, this dissertation constructs proper preconditioners from prior-knowledge of clutter to improve the performance of conventional STAP methods. The research involves the structure of the preconditioned process, the construction of STAP preconditioners and the combination of preconditioned method with traditional STAP methods. Simulations of SBR clutter mitigation verify the effectiveness of the preconditioned STAP method.Due to the earth's rotation or non-sidelooking radar configuration, the spectrum of SBR clutter varies with the range and shows non-stability. The non-stability of clutter will degrade STAP performance significantly and should be compensated. To compensate this non-stability of clutter, a new spectrum registration based method is proposed which uses non-uniformed discrete frequency sampling points. Specially, the mathematical model of the non-uniformed spectrum registration method is built. The choices of these non-uniformed discrete frequency samples and the estimate of clutter covariance matrix after compensation are also studied. Simulations show that the proposed spectrum registration method can compensate the non-stability effectively and achieve approximately optimal performance on condition that the distributing position of current clutter spectrum is given with high precision.Finally, since the space-time adaptive processors are usually realized on subarrays other than array elements, subarray partitions and corresponding weights will influence STAP performance. In this dissertation, the subarray STAP problem is solved separately in three cases. For the uniformed subarray partition case, the calculation formulas for optimal subarray weights are given with theoretical analysis. For the continually non-uniformed case, a serial particle swarm optimization (PSO) method is designed. The serial PSO method can deal with integer and real variables which correspond to the element numbers and weights of every subarray respectively, and gives a fast convergence rate as well. For the most complex case of non-continue partition, a genetic binary multiple PSO algorithm is proposed, which combines the advantages of the genetic algorithm and the PSO algorithm. Simulations and analysis are given for these three cases.