The Effect of Update Randomization on REVGEN Output: Broadband Energy Detector.
Abstract
Current advancements in computer simulation of ocean reverberation include the realtime implementation of REVEGEN(Reverberation Generator). The REVGEN algorithm which determines the randomization and number of update components can significantly affect the form of the reverberation return and correspondingly alter the false alarm and detection statistics for the response of a broadband energy detector. When the updates are limited and assigned constant amplitudes, there is a reduction in the probability of high intensity returns which, in turn, affect detector performance in two ways: (1) For a given signal-to-noise ratio, a signal produced with updates of constant magnitude will result in erroneously low probability of false alarm (PFA); and (2) These lowered PFA rates cause incorrect placement of detection thresholds which then yield erroneously high probabilities of detection. Both theoretical calculations and computer simulation show that increasing the number of updates resolves the statistical disparity. However, the reverberation algorithm which randomizes the update amplitudes according to a Rayleigh distribution is seen to yield the proper statistics independent of the number of updates used in the simulation. This decoupling of the detection statistics from the update algorithm may, for some simulation applications, provide the opportunity to reduce REVGEN computational requirements. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 31, 1978
- Accession Number
- ADA060121
Entities
People
- J. G. Melville
- M. E. Stegman