Simplified performance comparison metric based on asymptotic threshold ranking for MIMO radar estimation

Abstract

For many nonlinear estimation problems, classical lower bounds such as the Cramer-Rao bound (CRB) can characterize the mean squared error (MSE) performance only in the asymptotic region. While more powerful bounds like the Ziv-Zakai bound (ZZB) can also predict the best MSE performance in the nonasymptotic region, they may complicate the computation to an unaffordable extent. In this paper, for estimators with well-defined asymptotic threshold and virtually identical CRB, we propose to compare the MSE performance using the ZZB-based asymptotic threshold ranking. This method complements the CRB and is easier to compute than the ZZB, providing an efficient tool for preliminary system design.