Research On Robust Space-Time Adaptive Processing For Airborne Radar

Ground moving target indication(GMTI)system can realize the continual observation and tracing of ground moving targets in the key areas all over the world and further provide information for the location and imaging processing.Thus,it has drawn great attentions and been widely used in the military and civil tasks,e.g.,supervision of battlefield,scouting and striking,global observation and traffic controls.It has become one of the radar systems developed competitively by many countries.Under the ideal conditions that the priori knowledge of target steering vector and interference covariance could be accurately obtained,the space-time adaptive processing(STAP)system of airborne radars can provide a relatively good resolution and clutter suppression performance to realize the GMTI effectively.However,in the complex and varying practical environment,it is difficult for traditional STAP technologies to realize a satisfactory performanced.The quick and real-time GMTI processing under the conditions with few samples,inhomogeneous environment and existence of system errors is an important research field of robust STAP technology.It is significant for adapting the complex war environment and providing the accurate information of the moving target to commanders in the battlefield.Based on the actual requirements of GMTI application and the key research directions of the airborne STAP technology,space-time signal modeling,system performance analysis,error modeling and robust STAP technology are studied in this dissertation.The main contents and innovations are summarized as follows:The basic principles of STAP technology are studied in chapter ?.First,a modeling is conducted for the space-time signal model of airborne STAP system,and the characteristics of space-time power spectrum are briefly analyzed.It is pointed out that the clutter is coupled in space-time dimension.Thus,the clutter suppression should be jointly achieved from the space-time dimension.Then,the design criteria of some common STAP filters are discussed,and it is pointed out that the weight vectors of STAP obtained by these optimum criteria are equivalent.Next,some different algorithms for solving the optimum weight are given.Finally,in combination with the actual working conditions of the airborne radar,the common factors that deteriorate the performance of STAP under non-ideal conditions are stated,which establishes the theoretical basis for the researches in the following texts.The direct data domain(D3)STAP technology based on the steering vector estimation approach are studied in chapter ?.First,the basic principles of D3-STAP technology is studied.Due to the drawbacks of conventional single-sample D3-STAP technology,the multi-sample D3-STAP technology is proposed.Then,the common factors that deteriorate the performance of D3-STAP under non-ideal conditions are stated.Next,based on the above analysis,the D3-STAP technology based on the steering vector estimation approach is proposed.The real steering vector of the moving target is estimated by this technology,and the target-free training samples are obtained by using the estimated steering vector.Finally,the results of simulations verify the robustness of the proposed technology.The D-STAP technology based on the uncertainty set constraint is studied in chapter ?.First,the D-STAP technology based on the norm constraint is proposed,the optimal weight vector is solved by Lagrange multipliers approach,and it is pointed out that the D-STAP technology based on the norm constraint is equivalent to the D3-STAP technology based on a special diagonal loading approach.Then,according to the above analysis,the D3-STAP technologies based on the probability and worstcase constraint are derived,and the relationship between the uncertainty set constraint and the norm constraint is studied.Then,it is concluded that the D3-STAP technology based on the uncertainty set constraint should be superior to that based on the norm constraint.Finally,the simulation experiments verify the robustness of technologies proposed in this chapter.A fast l1-regularized STAP technology based on the alternating direction method of multipliers(ADMM)is studied in chapter ?.First,two commonly used l1-regularized STAP technologies,i.e.,l1-regularized recursive least-squares STAP and online coordinate descent STAP technologies,are studied.Then,the basic principles of ADMM approach are investigated.Next,in combination with the l1-regularized STAP application,a fast l1-ADMM-STAP technology is proposed,and the convergence rate and computational complexity of l1-ADMM-STAP technology are analyzed.Finally,by theoretical analysis and experimental verification,we show that the proposed algorithm provides a better output signal-to-clutter-noise ratio performance than other algorithms.The results of simulations approve the correctness of the derivations and analysis in this dissertation,and confirm the effectiveness of the advanced processing algorithms proposed.Thus,powerful supports for the design of airborne STAP algorithms and data processing in the case of complex environments are provided.