Accession Number : AD0763747

Title :   Detecting Signals in Poisson Noise.

Descriptive Note : Technical rept.,

Corporate Author : PRINCETON UNIV N J DEPT OF STATISTICS

Personal Author(s) : Watson,G. S.

Report Date : JUN 1973

Pagination or Media Count : 17

Abstract : It was shown that methods developed for tests of uniformity of directions in two dimensions can be usefully applied to detecting arbitrarily placed signals in Poisson noise, whether counts or times are given. Attention has been restricted to locally most powerful (invariant) tests because the distribution theory is available. Most powerful invariant distant tests may be derived for these situations but the distribution theory is lacking (Watson (1972)). The local tests seem more appropriate since greater care is required to detect small signals than large signals. The methods will be hardest to apply for signals of short duration. No progress was made for a signal whose shape is one of many.

Descriptors :   (*INFORMATION THEORY, *STATISTICAL ANALYSIS), TIME SERIES ANALYSIS, SAMPLING, NOISE, MEASURE THEORY, AERIAL RECONNAISSANCE, URANIUM, GEOPHYSICAL PROSPECTING, NUCLEAR RADIATION, DETECTION

Subject Categories : Statistics and Probability
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE