THEORY OF SIGNAL DETECTABILITY: COMPOSITE DEFERRED DECISION THEORY.
Abstract
The theory of sequential detection problems is extended to procedures of a composite signal hypothesis, such as signals whose amplitude or phase is unknown. Observations of signal-and-noise are assumed not to be independent so that information can be extracted about the signal parameters for use in the detection process. The yes-no terminal decision must be made within a finite period. A general theory is presented for the solution of practical problems in the sequential detection of composite signal hypotheses. The theory specifies the optimum stopping rule needed for the sequential detector and the dependence between the observer's opinion of the cause of the reception and the distribution of the unknown signal parameters. The general theory implies the information that must be extracted from the observation is the likelihood ratio and the a posteriori signal parameter distribution. The form of the optimum detector, derived from the general theory, must include an adaptive capability, so that the signal parameter distribution can be sequentially updated. Applications of the general theory to a signal of unknown phase and a signal of unknown amplitude are presented. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1965
- Accession Number
- AD0614480
Entities
People
- Richard A. Roberts
Organizations
- University of Michigan