DETECTION OF CW SIGNALS IN NOISE AND COMPARISON WITH PULSED SIGNAL DETECTION: EXPERIMENTAL STUDY,
JOHNS HOPKINS UNIV SILVER SPRING MD APPLIED PHYSICS LAB
Pagination or Media Count:
With the noise set to a convenient value and no CW signal present, the false alarm probabilities as a function of threshold setting were recorded. Signal was then added to give a desired signalnoise ratio, and the detection probabilities were plotted as a function of threshold setting. This was repeated for several signal-to-noise ratios to give a family of curves corresponding to the video bandwidth used. The video bandwidth was then changed and the entire procedure repeated. In all, four bandwidths were used corresponding to ratios of i-f half-bandwidthvideo bandwidths of 1, 3, 10 and 30. The data have been plotted in Figures 2, 3, 4 and 5. The threshold parameter has been eliminated by plotting the detection probabilities for a given threshold against false alarm probabilities corresponding to that threshold. In Figure 6, signal-to-noise ratio versus probability of detection is plotted for a constant false alarm probability of 12000. The Marcum theoretical curves for the pulse case are plotted for comparison with the experimental CW data.