I. Introduction
Performance bounds on parameter estimation are of utmost importance in statistical signal processing theory and applications. In addition to providing a benchmark for performance analysis of estimators [1], [2], [3], [4], they are frequently used for system design and feasibility study (see e.g. [5], [6], [7], [8], [9]) and in the recent years, they are employed as an optimization criterion in sequential cognitive systems [10], [11], [12], [13], [14], [15], [16], [17], in both the Bayesian and non-Bayesian (frequentist) approaches. The quality and relevance of the bounds are determined by the following factors: 1) tightness and attainability, 2) computational complexity, and 3) ability to provide insight into the considered problem.