Detection in Non-Gaussian Background,
Abstract
The problem of detecting a small signal in a non-Gaussian background is discussed analytically for a very limited set of conditions. It is supposed that the background contains bursts of noise whose complete rejection might aid detection. For several specific detection procedures, formulas are derived by which empirically known distribution functions of the background may be used to predict a change in detection threshold resulting from burst rejection. Particular attention is given to the use of the first-order distribution function, even though it alone does not completely determine the results. A suggestion by Birdsall connecting this study with the use of the likelihood ratio is also discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 02, 1958
- Accession Number
- ADA053117
Entities
People
- Philip Rudnick
Organizations
- Scripps Institution of Oceanography