A Unified Theory of Adaptive Subspace Detection Part II: Numerical Examples

This paper is devoted to the performance analysis of detectors proposed in companion [1] where a comprehensive design framework presented for adaptive detection subspace signals. The addresses four variations on detection: may be known or only by its dimension; consecutive visits unconstrained they constrained prior probability distribution. In this paper, Monte Carlo simulations are used compare derived with estimate-and-plug (EP) approximations generalized likelihood ratio (GLR) detectors. Remarkably, some EP appear here first time (at least best authors' knowledge). numerical examples indicate that GLR effective partially-known signals affected inherent uncertainties due system operating environment. particular, if signal known, tend ouperform If, instead, dimension, and very similar. Actually, there does not exist general rule recommending first-order approach respect second-order one vice versa. Nevertheless, contains specific case can outperform schemes.