Multi-Channel Fast Parametric Algorithms and Performance for Adaptive Radar

Airborne radar systems employing radar sensor arrays utilize multi-channel (MC) signal processing techniques for optimal detection and localization of targets. The detection and localization statistics have mathematical structures that typically require the inverse of an estimated covariance matrix in order to be evaluated. Due to the size of sensor arrays and the number of pulses in a coherent processing interval (CPI), the dimension of the covariance arrays is very large (1000s) and the computational burden of estimating and inverting such large arrays has led to the development of parametric methodologies that significantly reduce both the computational requirements and the amount of measured data to create the inverse co-variance matrix estimate. Recent developments have indicated non-stationary covariance estimates may provide enhanced detection performance over stationary covariance estimates. This paper outlines the fast computational structure possibilities of both stationary and non-stationary covariance inverse structures in one-dimensional and multi-channel cases.