Accession Number : AD0769497

Title :   Multiple-Two-Hypothesis Testing - A New Method for Determining Signal Content.

Descriptive Note : Master's thesis,

Corporate Author : AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OHIO SCHOOL OF ENGINEERING

Personal Author(s) : Nowacki,John A.

Report Date : SEP 1973

Pagination or Media Count : 122

Abstract : Detecting and classifying overlapping waveforms can be done using M-ary hypothesis testing. A new detection method, multiple-two-hypothesis testing, is developed by minimizing the probability of error for each decision made, and is shown to have computational advantages over M-ary hypothesis testing. A sequential test is developed from multiple-two-hypothesis testing, and it is also shown to have computational advantages over M-ary hypothesis testing. The three detection methods are applied to the synchronous detection of two signals with no overlap, and their performances are compared. Then M-ary hypothesis testing and multiple-two-hypothesis testing are applied to the synchronous detection of two signals with one overlap and error equations are presented for the two methods. Finally, multiple-two-hypothesis testing is combined with maximum a posteriori estimation to solve nonsynchronous detection problems. (Author)

Descriptors :   (*Information theory, Statistical analysis), (*Signal processing, White noise), (*Waveforms, Classification), Hypotheses, Decision theory, Probability density functions, Statistical tests, Sonar, Radio astronomy, Radar, Theses

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE