ON RELATIONSHIPS BETWEEN DETECTION AND ESTIMATION THEORY.
This paper takes a global approach to the processing of information. The viewpoint is Bayesian where any uncertain parameters are modeled as random variables and knowledge about them is summarized by a priori probability density functions. The processors discussed must decide if the random processes observed consist of a signal obscured by noise or noise alone. The resulting detectors are optimum in the sense of making a least risk decision. The general form of the likelihood ratio is derived. Research results from the area of array processing are presented.