Fast Block LMS and RLS-Based Parameter Estimation and Two-Dimensional Imaging in Monostatic MIMO RADAR Systems With Multiple Mobile Targets

Abstract

This paper presents novel adaptive schemes and the pertinent analysis for estimation of the number of targets, their associated radar cross section (RCS) values and Doppler velocities in monostatic MIMO radar systems. These schemes are based on the fast block least mean squares (FBLMS) and fast block recursive least squares (FBRLS) algorithms considering both stationary as well as mobile targets and/or radar platform. For the stationary scenario, schemes are proposed for estimation of RCS coefficients, target number and their angle, range locations. In addition, a procedure is also developed to obtain the 2D-image of the MIMO radar scene in angle-range dimensions. Analysis is presented for the first and second order moments of the observation as well as estimation errors resulting from the proposed FBLMS, FBRLS techniques, along with their global convergence. Both the proposed schemes are shown to converge globally to the optimal Wiener filter, which is lacking in existing schemes. Further, a Bayesian information criterion based selection rule is employed to enhance the estimation and imaging performance of the proposed adaptive frameworks. Cramer-Rao bounds are derived to characterize the mean squared error of the estimated RCS coefficients for stationary scenarios and also for joint RCS, Doppler velocity estimation in mobile MIMO radar scenarios. The FBLMS and FBRLS schemes are shown to have significantly lower computational complexities in comparison to existing MIMO radar schemes. Simulation results demonstrate the improved performance of the proposed schemes and also validate the derived analytical expressions.