Two-Sample Non-parametric Detectors Using Dependent Samples

✍ Scribed by Emad Al-Hussaini; Farid Badran; Laurence Turner

• Publisher: Elsevier Science
• Year: 1977
• Tongue: English
• Weight: 804 KB
• Volume: 304
• Category: Article
• ISSN: 0016-0032
• DOI: 10.1016/0016-0032(77)90006-0

✦ Synopsis

perfomtance of two-sample non-paramedic detectors using dependent samples is considered. It is shown that in the situation in which square-law envelope detection is employed, the signals to be detected fluctuate in accordance with a chi-square distibution and the background interference is additive Gaussian noise; the two-sample Generalized Sign (GS), Mann-Whitney (MW), Modified Savage (MS) and Modified Rank Square (MRS) detectors have superior asymptotic processing times to those that are obtained when independent samples are used. Also, it is shown that in addition to the advantage of improved performance, detection using dependent samples can have implementation advantages as compared with detection based on the use of independent samples.

* T-dependent means that samples taken T or more set apart are statistically independent, and that if they are taken less than T set apart then they are statistically dependent.