Rank rule for signal detection in noise of unknown level
Abstract
A locally most powerful rank rule is derived for signal detection against a noise background of unknown intensity. The rule is shown to be applicable to incoherent detection of both nonfluctuating and fluctuating signals, including signals that fluctuate according to the Rayleigh-Rice law. The rule possesses a parametric property in that it ensures an invariable false-alarm probability for any noise level or noise distribution, without requiring a change in threshold, and in that it maximizes the detection probability at small signal to noise ratios.
- Publication:
-
Radioehlektronika
- Pub Date:
- January 1975
- Bibcode:
- 1975Radel..18..121P
- Keywords:
-
- Background Noise;
- Ranking;
- Signal Detection;
- Random Processes;
- Rayleigh Distribution;
- Signal To Noise Ratios;
- Communications and Radar