Detection of random transient signals via hyperparameter estimation

Difficulties arise with the generalized likelihood ratio test (GLRT) in sit uations where one or more of the unknown signal parameters requires an enum eration that is computationally intractable. In the transient signal detect ion problem, the frequency characteristics of the signal are typically unkn own; therefore, even if an aggregate signal bandwidth is assumed, the estim ation problem intrinsic to the GLRT requires an enumeration of all possible sets of signal locations within the monitored band. In this paper, a prior distribution is imposed over those portions of the s ignal parameter space that traditionally require enumeration. By replacing intractable enumeration over possible signal characteristics with an a prio ri signal distribution and by estimating the "hyperparameters" (of the prio r distribution) jointly with other signal parameters, it is possible to obt ain a new formulation of the GLRT that avoids enumeration and is computatio nally feasible, The GLRT philosophy is not changed by this approach-what is different from the original GLRT is the underlying signal model. The perfo rmance of this new approach appears to be competitive with that of a scheme of emerging acceptance: the "power-law" detector.