A COMPARISON OF THREE MODELS FOR SEQUENTIAL OBSERVATIONS IN SONAR DETECTION.
Abstract
The paper compares three mathematical models which have been proposed to describe the sonar detection problem in the case of sequential observations. The first model assumes complete independence between observations. The second model assumes a constant dependence between observations, independent of the time between observations. Finally, the third model assumes a dependence between observations which decays exponentially as the time between observations increases. These models have been presented in a consistent and unified manner to facilitate comparison. As would be expected, the independent model requires uniformly smaller signal-to-noise ratios for any probability of detection than the other two models and the third model requires uniformly smaller signal-to-noise rations than does the second model. However, in the range of detection probabilities 0.3-0.7, the maximum increase in signal-to-noise ratio required for the constant correlation model over the exponentially decaying model is less than 14 percent. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1966
- Accession Number
- AD0635394
Entities
People
- Donald Iglehart
Organizations
- Institute for Defense Analyses