The Detection of Time-Varying Laser Signals.
Abstract
A sequential detector is developed for low-intensity, time-varying laser signals. The performance of this detector is superior to that of any fixed-sample-size detector designed for the same purpose. It is shown that all previously-developed detectors are special cases of the sequential detector. The laser signals are first pre-detected by a wide-bandwidth photomultiplier, assumed to be able to resolve the times-of-emissions of the photoelectrons. The joint probability distribution function is derived for these times-of-emissions, which form a non-stationary Poisson process. A truncated sequential probability ratio test for this non-stationary process is devised; it is shown to be a generalization of all previously-developed tests for Poisson processes, both sequential and fixed-sample-size. Fundamental differences between this test and those for signals in Gaussian noise are described. A method using stochastic matrices is developed to evaluate the performance of this time-varying sequential test. In many cases, the probabilities of making decisions and the average time to termination can be found exactly. In other cases, no analytical solution is possible; thus the performance of the test is evaluated by a digital-computer simulation. Several practical examples are given to illustrate the advantages to be gained in decision time when the sequential detector is used for laser signals. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1970
- Accession Number
- AD0879622
Entities
People
- Kenneth Jay Lutz
Organizations
- Johns Hopkins University