(The) performance comparison of the STAP algorithms for slow moving ground target detection

Abstract

In airborn radar system the ground clutter is the most severe interference which must be suppressed to detect moving target. Space-Time Adaptive Processing(STAP) is the extended version of the general adaptive array processing in two dimension. STAP combines signals received from multiple antenna element and from multiple pulses in CPI. The complexity problem and data support for covariance matrix estimation motivates partially adaptive STAP.The non-adaptive preprocessor for reducing the dimension is used. The four types of preprocessor are discussed and the performance is compared to the optimal case. Another issue is a non-homogeneous interference environment which results in performance degradation. The non-homogeneous environment induces two problem in STAP. One is the data support problem in covariance matrix estimation. The other is the existence of discrete non-homogeneities in primary range cell. For the former, the Non-Homogeneous-Detection is used for selecting homogenous secondary data for good estimation of covariance matrix. To solve the latter problem, the $D^{3}$ algorithm is considered. As a result, the hybrid algorithm which combines advantages of both JDL and $D^{3}$ algorithm is discussed. The performance of the JDL only and hybrid algorithm was compared in non-homogeneous environment.