Performance Improvement of Subspace-Based Direction-Finding Algorithms Using Higher-Order Statistics

摘要

Traditional array signal processing techniques have been relying on the use of received signal's second-order statistic for many years. However, it suffers with some fundamental limitations. Studies of array processing based on higher-order statistic has been proposed aiming to overcome these limitations. This paper is aimed to assess the array performance enhancement when using higher-order statistic from the differential geometry perspective. Defined as the locus of all array response vectors over a set of signal parameters, the array manifold's geometrical shape and properties are known to be crucially important in characterizing the array performance. In this paper, the geometry of an array manifold associated with a higher-order statistic is investigated using of the concept of virtual sensor array. Performance analysis is presented to examine the array performance enhancement both in terms of the Cramer Rao lower bound and the array detection and resolution capabilities.