SIGNAL-DETECTION IN MULTIVARIATE CLASS-A INTERFERENCE

The aim of this article is to identify and evaluate multivariate density functions when the univariate densities are all Class-A. Three models are proposed: Model A for independent random variables; Model B for dependent random variables; Model C for uncorrelated random variables. For each model, the structure of the marginal densities, moments, and characteristic functions are evaluated. Truncated forms for the multivariate densities are proposed. The error and percentage error due to truncation are evaluated numerically for the two-dimensional case. Finally, a signal detection problem is examined. The performance of the optimal detector is compared to that achieved by the detector designed using truncated density functions. It is shown that the performance difference Is negligible when the model parameters are small.