ELF Noise Statistics and Processing under Disturbed Conditions.
Abstract
This report addresses the question of whether nonlinear processing schemes optimized on the basis of ambient extremely-low-frequency (ELF) atmospheric noise data will give near-optimum processing gain in nuclear/PCA environments. Analytic expressions are derived for noise-pulse waveforms, amplitude probability distributions (APDs), and processing gain as a function of clip levels. Numerical results are given for each of these quantities under ambient and spread-debris-nuclear/PCA environments. Good agreement is obtained with ambient ELF atmospheric noise data and results from much more complicated numerical treatments. The ionospheric disturbances considered significantly increase the ratio of the energy carried in noise spikes to the energy of the background Gaussian noise. This change in ratio, which alters the APDs considerably, occurs because Gaussian noise propagates over longer distances than noise spikes, and therefore tends to be suppressed relative to the spikes by increased attenuation. On the other hand, total RMS noise, to which noise spikes make the major contribution, may well be affected only slightly by environmental changes. Consequently, to achieve maximum processing gain, clip levels should be set according to the ratio of noise-spike energy to Gaussian background energy. Because this ratio can change drastically, an adaptive procedure is called for. Calculations show that the experimental SANGUINE noise-suppression circuit, which self-adapts to clip some specified fraction--typically 20 to 60 percent--of the time gives nearly optimum processing gain for a wide variety of ambient and nuclear/PCA environments.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1974
- Accession Number
- ADA042116
Entities
People
- E. C. Field
- R. Berlot