Fast And Robust Space-time Adaptive Processing Method For Nonhomogeneous Environment

Space-time adaptive processing(STAP)can effectively suppress the clutter and achieve better moving target detection by the joint adaptive filtering in spatial-temporal domain for airborne phased array radar,and it has become a key technique of modern airborne phased array radar.However,the complex and volatile working scenarios of airborne phased array radar bring out sever challenge to STAP,it is urgent and necessary to develop fast and robust STAP methods for the practical complex clutter environment.Therefore,by employing the theoretical analysis and verification of actual measured airborne phased array radar data,this dissertation meets closely around the requirement of clutter suppression in practical complex clutter environment,and mainly investigates the fast and robust STAP methods with small training sample support,fast calculation methods for STAP weight vectors and the nonhomogenous clutter suppression methods.The related research in this dissertation can effectively improve the convergence and robustness for clutter suppression in practical complex clutter environment.The detailed contents and main conclusions can be summarized as follows:The second chapter introduces the signal model of received space-time data for airborne phased array radar.Moreover,the intrinsic sparsity of STAP is analyzed from the view of the clutter spatial-temporal power spectrum and the space-time adaptive weight vectors,which provides a basis for the following researches.The third chapter studies the clutter suppression methods with small training sample support.As it is known that the number of homogeneous training samples is insufficient in practical environment,therefore the clutter suppression of conventional STAP methods deteriorates significantly.In order to solve this problem,based on the sparsity of clutter spatial-temporal power spectrum,a robust and fast iterative sparse recovery method for STAP is firstly proposed in this chapter.In the proposed method,the sparse recovery of clutter spatial-temporal spectrum and the calibration of space-time overcomplete dictionary are achieved iteratively.The proposed method can not only alleviate the effect of noise and dictionary mismatch,but also reduce the computational cost by recursive inverse matrix calculation,which is due to direct matrix inversion.Based on the simulated and the actual airborne phased array radar data,it has been verified that the proposed method can provide better performance with small training sample support in practical complex non-homogeneous environment.Afterwards,based on the sparsity of the space-time adaptive weight vectors,a fast STAP method based on projection approximation subspace tracking(PAST)with sparse constraint is proposed in this chapter.In the proposed method,based on the low-rank property of the clutter covariance matrix,a sparse constraint is imposed in the cost function of PAST,and the adaptive weight vector is then derived iteratively.Because of the sparse constraint in PAST,the proposed method provides a more robust and stable estimation of the clutter subspace when only a small set of training samples is available.Results using simulated and actual airborne phased array radar data verify the effectiveness of the proposed method.The fourth chapter studies the fast calculation methods for STAP weight vectors.Because of large computational complexity in the inverse space-time covariance matrix computation,the STAP methods cannot satisfy the real-time requirement for practical implementation.In order to solve this problem,Based on the block Hermitian matrix property of space-time covariance matrix,a new element-order recursive method is proposed in this chapter to calculate the inverse space-time covariance matrix for STAP weight vector.In the proposed method,the inverse space-time covariance matrix of first element-order is initially calculated recursively based on block Hermitian matrix property,and then the inverse space-time covariance matrix of high element-order is correspondingly deduced recursively based on obtained inverse covariance matrix of previous element-order.Finally,STAP weight vector is calculated based on the final inverse covariance matrix.Moreover,a modified reduced-dimension STAP method is derived by combing the proposed method with the reduced-dimension STAP method for more computational cost saving.Based on the simulated and the actual airborne phased array radar data,the clutter suppression speed can be effectively improved by utilizing the proposed method.The fifth chapter studies the non-homogenous discrete interference suppression methods.It is known that the conventional STAP methods would suffer severely performance loss because the discrete interference is not contained in the training data for clutter covariance matrix estimation.Meanwhile the Direct data domain(DDD)method only utilizes the CUT data,the statistical property about the homogeneous clutter is ignored,and therefore the DDD method cannot suppress the homogeneous clutter effectively.Moreover,the hybrid discrete interference suppression methods based on adaptive localized domain transformation(DDD-ALT)cannot provide desirable suppression performance when sufficient number of I.I.D.training data is collected.In order to solve this problem,by exploiting the intrinsic sparsity of the clutter and discrete interference in the angle-Doppler domain,a space-time discrete interference suppression method based on robust Bayesian compressive sensing(BCS)is proposed in this chapter.In the proposed method,the estimation of the spectral distribution and the calibration of overcomplete space-time dictionary are achieved iteratively based on the fast BCS and the cost function minimization.The robust sparse recovery of the clutter and discrete interference profiles can be achieved at low computational complexity.Results utilizing simulated and actual airborne phased array radar data verify the effectiveness of the proposed method in nonhomogeneous environment.The sixth chapter studies the non-homogeneous interference target signals(outliers)methods.The interference-targets(outliers)are inevitable contained in the training samples set,thus the clutter covariance matrix cannot be estimated accurately,leading to degraded STAP performance.The conventional non-homogeneity detector based on genral inner product(GIP)and prolate spheroidal wave functions(PSWF)directly removes training samples with outliers,when the environment is considerably non-homogeneous,large numbers of training samples would be removed,thus the clutter covariance matrix cannot be well estimated for desirable clutter suppression as the number of training samples is deficient.In order to solve this problem,a robust non-homogeneity clutter suppression method based on knowledge aided sparse recovery is proposed in this chapter.In the proposed method,the spectral profiles of the clutter and outliers are firstly estimated by sparse recovery processing.Then based on the system prior parameters,the clutter mask is constructed to select the space–time steering vectors corresponding to the clutter and outliers components.Afterwards the clutter suppression is achieved based on the clutter subspace obtained from the selected space–time steering vectors corresponding to the clutter.Because the clutter and outlier profiles are effectively estimated and distinguished by the knowledge aided sparse recovery processing,robust clutter subspace estimation can be achieved for clutter suppression,clutter subspace estimation is not affected by the number of homogeneous training samples.Through the simulated and actual airborne phased array radar data,it is verified that the proposed method can effectively improve the clutter suppression performance in nonhomogeneous outlier environment.The seventh chapter makes a summary of the dissertation,while several open research directions for the STAP methods are proposed.