Turbo Parametric Spectral Estimation Method of Clutter Profile for Adaptive Radar Detection

Identification of unwanted components of a received echo is critical to improving the radar detection performance. Previously proposed parametric parameter estimation methods, such as MUSIC, ESPRIT, and Burg were implemented to estimate moments of radar clutter. Since none of these methods could estimate the value of Doppler spread and have sufficient accuracy, the Stochastic Maximum Likelihood (SML) technique was implemented. Since its estimation accuracy was profoundly initial point dependent and computationally costly, a novel estimation method (Turbo SML) is proposed. The proposed method outperformed the strategies proposed in the literature with its high Doppler resolution, accuracy, and low computational complexity. Besides, Turbo SML performance was optimized by using Burg estimates for starting point choice. After accomplishing nearly optimal estimation, its estimates were utilized to implement an approximately Max-Normalized Signal to Interference plus Noise Ratio (SINR) filter. Superior to detection filters in literature, the proposed filter can maximize its output Signal to Interference plus Noise Ratio (SINR) with a few numbers of secondary data.