Low RCS Target Tracking in Estimated Rapidly Varying Sea Clutter Using a Kronecker Product Approximation Algorithm

Abstract

We propose a method for estimating the space-time covariance matrix of rapidly-varying sea clutter following a dynamic state space matrix model. The covariance matrix dimension can become computationally infeasible as it increases with the number of range bins and dwell pulses required for coherent processing. In order to reduce the computational complexity, we apply the Kronecker product (KP) approximation and particle filtering to estimate the space-time covariance matrix, and we demonstrate the proposed method's validity using real clutter data. We also demonstrate that the method ensures that the covariance matrix estimate is always positive definite. The covariance matrix estimation is integrated with a track-before-detect filter for tracking a low radar cross-section (RCS) target in strong sea clutter.