Monopulse processing for tracking unresolved targets

When target echoes interfere (i.e., the echoes are not resolved in the frequency or time domains) in a monopulse radar system, the Direction-Of-Arrival (DOA) estimate indicated by the in-phase monopulse ratio can wander far beyond the angular separation of the targets. In addition to closely-spaced targets, the problem of unresolved or merged measurements also occurs when targets are observed in the presence of jammer signals or sea-surface-induced multipath. The failure to detect the presence of this interference and address it in the DOA estimation can be catastrophic to the performance of the tracking algorithm, since its position and velocity estimates determine the association of any subsequent measurements to the target.;Monopulse processing for tracking unresolved targets is addressed in four parts. The first part involves the development of the Probability Density Function (PDF) and statistics of the measured amplitude of the sum signal for an arbitrary number of unresolved Rician targets. The PDF and statistics are utilized to develop estimators of the target amplitude parameters, which define the SNR of the target, and discriminators for models of the target amplitude fluctuations. The second part involves the development of the joint PDF and statistics of the complex monopulse ratio for an arbitrary number of unresolved Rician targets and a fixed-amplitude target in the presence of multipath. The information contained in the sum-channel signal is retained in the PDF of the complex monopulse ratio by conditioning the PDF on the measured amplitude of the sum signal. The third part involves DOA estimation for single resolved targets and two unresolved targets. DOA estimators are developed for two unresolved Rayleigh targets with known relative Radar Cross Section (RCS) and a Rayleigh target in the presence of a Gaussian jammer. Estimators of the variance of the DOA estimates are also developed so that tracking of the targets can be accomplished with a nonstationary filtering technique such as the Kalman filter. The fourth part involves the detection of the presence of unresolved targets. A Generalized Likelihood Ratio Test (GLRT) is used to develop a Neyman-Pearson algorithm for the detection of the presence of unresolved Rayleigh targets, and performance predictions of the new algorithm are shown to agree rather well with simulated performance. Cramer Rao Lower Bounds are developed for the various parameter estimators, and the performance of the estimators and detection algorithms are illustrated through the results of Monte Carlo simulations.